Alma Mater Studiorum Università di Bologna

MASTER’S DEGREE COURSE IN ADVANCED AUTOMOTIVE

ELECTRONIC ENGINEERING

Master’s Thesis in Automotive Electronic Engineering

SMART DATA ANALYSIS ON

LAMBORGHINI CONNECTED VEHICLES

Author: University Supervisor:

Davide Palombo Prof. Carlo Augusto Grazia

Company Supervisor:
Ing. Luca Giardino

Session October

A.Y. 2021/2022
La determinazione sta nel fare
delle cose che altri non fanno.
 Abstract

The thesis is the result of work conducted during a period of six months at the
Strategy department of Automobili Lamborghini S.p.A. in Sant’Agata Bolognese
(BO) and concerns the study and analysis of Big Data relating to Lamborghini’s
connected cars. The Big Data is a project of Connected Car Project House, that is
an inter-departmental team which works toward the definition of the Lamborghini
corporate connectivity strategy and its implementation in the product portfolio.
The Data of the connected cars is one of the hottest topics right now in the au-
tomotive industry; in fact, all the largest automotive companies are investi,ng a
lot in this direction, in order to derive the greatest advantages both from a purely
economic point of view, because from these data you can understand a lot the be-
haviors and habits of each driver, and from a technological point of view because it
will increasingly promote the development of 5G that will be an important enabler
for the future of connectivity.
The main purpose of the work by Lamborghini prospective is to analyze the data of
the connected cars, in particular a data-set referred to connected Huracans that had
been already placed on the market, and, starting from that point, derive valuable
Key Performance Indicators (KPIs) on which the company could partly base the
decisions to be made in the near future.
The key result that we have obtained at the end of this period was the creation of
a Dashboard, in which is possible to visualize many parameters and indicators both
related to driving habits and the use of the vehicle itself, which has brought great
insights on the huge potential and value that is present behind the study of these
data.
The final Demo of the project has received great interest, not only from the whole
strategy department but also from all the other business areas of Lamborghini,
making mostly a great awareness that this will be the road to follow in the coming
years.
Contents

1 Introduction 1
1.1 The Four Automotive Trends: CARE . . . . . . . . . . . . . . . . . . 2
1.1.1 Connected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.2 Autonomous . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.1.3 Redefined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1.4 Electrified . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 The Importance of Big Data 10

2.1 How to define success for my organization? . . . . . . . . . . . . . . . 11
2.2 Are laws threatening opportunities? . . . . . . . . . . . . . . . . . . . 12
2.2.1 Data Privacy Legislation . . . . . . . . . . . . . . . . . . . . . 12
2.2.2 Right to Repair . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2.3 European Commission’s Data Act . . . . . . . . . . . . . . . . 12
2.3 How to catch up with the competition? . . . . . . . . . . . . . . . . . 13
2.4 Who’s who in the automotive data industry . . . . . . . . . . . . . . 14
2.5 Data monetization opportunity . . . . . . . . . . . . . . . . . . . . . 17
2.6 What is Data Monetization and why are automakers trying to mon-
etize vehicle data? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.7 What does the future hold for automotive data? . . . . . . . . . . . . 23

3 Lamborghini Projects on Big Data 26

3.1 First PoC launched by Lamborghini . . . . . . . . . . . . . . . . . . . 27
3.2 Second PoC launched by Lamborghini . . . . . . . . . . . . . . . . . 29
3.3 Third PoC launched by Lamborghini . . . . . . . . . . . . . . . . . . 30
3.4 Fourth PoC launched by Lamborghini . . . . . . . . . . . . . . . . . . 33

4 Data Analysis and Dashboard Implementation 34

4.1 Data Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2.1 Route Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.2.2 Driving Mode Analysis . . . . . . . . . . . . . . . . . . . . . . 37
4.2.3 Circuit Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.2.4 Temporal and Frequency Analysis . . . . . . . . . . . . . . . . 42
4.2.5 Nearby Service Shop Analysis . . . . . . . . . . . . . . . . . . 42
4.2.6 Fleet Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2.7 Weather Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.2.8 Country Comparison Analysis . . . . . . . . . . . . . . . . . . 48

I
 4.3 Privacy and data protection in connected vehicles . . . . . . . . . . . 49
4.3.1 Privacy and data protection in connected vehicles: the principles 50
4.3.2 Privacy and data protection in connected vehicles: Regulations 50
4.3.3 How Privacy was addressed for PoC purposes . . . . . . . . . 51

5 Lamborghini KPI 53
5.1 Driving Mode KPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.2 Circuit KPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.3 Nearby Service Shop KPI . . . . . . . . . . . . . . . . . . . . . . . . 58
5.4 Weather KPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
5.5 Country Comparison KPI . . . . . . . . . . . . . . . . . . . . . . . . 61
5.6 Electrification KPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

6 Next Steps 63
6.1 Next Steps: Short Term . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.2 Next Steps: Middle Term . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.3 Next Steps: Long Term . . . . . . . . . . . . . . . . . . . . . . . . . . 65

7 Conclusion 67

II
List of Figures

1.1 Percentage of consumers who feel that increased vehicle connectivity

will be beneficial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 By 2030, 45 percent of global new-car sales could be at level 3 or
above in connectivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Percentage of consumers who agree that autonomous vehicles will not
be safe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 While the challenges are significant, they in turn provide great op-
portunities for players to conquer new markets and further reduce
costs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Consumer powertrain preferences for their next vehicle (2020) . . . . 6
1.6 Reasons consumers consider hybrids or BEVs . . . . . . . . . . . . . 7
1.7 Consumers considerations about BEV/PHEV vehicles . . . . . . . . . 9

2.1 Data Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.2 Automakers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 Drivers affecting connected car data . . . . . . . . . . . . . . . . . . . 16
2.4 Barriers affecting connected car data . . . . . . . . . . . . . . . . . . 16
2.5 Data ecosystem overview . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.6 Why automakers are trying to monetize vehicle data . . . . . . . . . 20
2.7 Internal Data monetization . . . . . . . . . . . . . . . . . . . . . . . . 20
2.8 External Data monetization . . . . . . . . . . . . . . . . . . . . . . . 21
2.9 How the current vehicle data monetization ecosystem looks like . . . 22
2.10 Data volumes and costs could skyrocket by 2025 . . . . . . . . . . . . 24

3.1 Lamborghini cluster customization . . . . . . . . . . . . . . . . . . . 28

3.2 Visualization of the clusters . . . . . . . . . . . . . . . . . . . . . . . 28
3.3 Big Data as enabler for the electrification . . . . . . . . . . . . . . . . 29
3.4 Statistical data analysis on Dashboard . . . . . . . . . . . . . . . . . 31
3.5 Dashboard for digital marketing . . . . . . . . . . . . . . . . . . . . . 32
3.6 Connected services usage Dashboard . . . . . . . . . . . . . . . . . . 32
3.7 Data Lake architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.1 Ecosystem of the Architecture . . . . . . . . . . . . . . . . . . . . . . 34

4.2 Data Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.3 Road and places most frequented . . . . . . . . . . . . . . . . . . . . 37
4.4 Driving mode usage by country . . . . . . . . . . . . . . . . . . . . . 38
4.5 Driving mode clusters per model . . . . . . . . . . . . . . . . . . . . . 38
4.6 Driving mode usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

III
4.7 Most visited circuits in Italy . . . . . . . . . . . . . . . . . . . . . . . 40
4.8 Number of visits to circuits by month . . . . . . . . . . . . . . . . . . 40
4.9 How long does it take between a visit to a circuit and another? . . . . 41
4.10 Number of unique VIN to the circuit for each visit number . . . . . . 41
4.11 Monthly and weekly usage of cars . . . . . . . . . . . . . . . . . . . . 42
4.12 At what time a journey usually begins? . . . . . . . . . . . . . . . . . 42
4.13 Total number of visits per service shops . . . . . . . . . . . . . . . . . 43
4.14 Number of visits to service shops by time . . . . . . . . . . . . . . . . 44
4.15 How long does it take between a visit to the service shop and another? 44
4.16 The general values attributed to the fleet . . . . . . . . . . . . . . . . 45
4.17 Number of active connected cars per moth . . . . . . . . . . . . . . . 45
4.18 Tons of CO2 produced per month . . . . . . . . . . . . . . . . . . . . 46
4.19 Weather conditions of use of the vehicle by city . . . . . . . . . . . . 47
4.20 Driving mode distribution by weather conditions . . . . . . . . . . . . 47
4.21 Vehicle usage divided by country . . . . . . . . . . . . . . . . . . . . 48
4.22 Driving mode usage by country . . . . . . . . . . . . . . . . . . . . . 48
4.23 At what time a journey usually begins? . . . . . . . . . . . . . . . . . 49

5.1 Driving mode clusters could be used as KPI . . . . . . . . . . . . . . 54

5.2 Driving mode variation clusters could be used as KPI . . . . . . . . . 54
5.3 Driving mode variation by country could be used as KPI . . . . . . . 55
5.4 Number of unique VIN to the circuit used as KPI . . . . . . . . . . . 57
5.5 Relation between driving mode most used by the vehicles that have
been seen near a service shop as KPI . . . . . . . . . . . . . . . . . . 59
5.6 Relation between the usage of the vehicle and the weather condition
as KPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.7 Country comparison among at what time a journey usually begins as
KPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.8 Average duration of a journey used as KPI . . . . . . . . . . . . . . . 62

IV
Chapter 1

Introduction

Since the introduction of the smartphone, it has become clear that customers are
quick to adopt even highly complex and expensive technology if it makes their lives
easier. In other words, users value convenience and ease. These core values turned
the automobile into the defining technical cultural item of the 20th century. Now
it is time to translate these properties into the context of today’s – and tomorrow’s
– technology and society. The automotive industry has the opportunity to shape
this fundamental restructuring. When devising strategies and business models, com-
panies should not only consider direct product purchasers but all users and groups
affected by transport issues. The automobile has long since changed from a technical
to a social commodity: it guarantees our personal mobility and social participation,
shapes our cities and landscapes, and structures our temporal and spatial thinking.
This is why we have to rethink the whole automotive industry – with the focus on
the use rather than the production of vehicles, in order to make the lives of individ-
ual users more enjoyable, more efficient and safer.
Furthermore, today’s economies are dramatically changing, triggered by develop-
ment in emerging markets, the accelerated rise of new technologies, sustainability
policies, and changing consumer preferences around ownership. Digitization and
new business models have revolutionized other industries, and automotive will be
no exception. For the automotive sector, these forces are giving rise to four disrup-
tive technology-driven trends:

• Connected

• Autonomous

• Redefined

• Electrified

Most industry players and experts agree that these four trends will reinforce and
accelerate one another, and there is general consensus that the industry is ripe for
disruption. Yet although the widespread sentiment that game-changing disruption is
already on the horizon, there is still no integrated perspective on how the automotive
industry will look in 10 to 15 years as a result of these trends.

1
1.1 The Four Automotive Trends: CARE
The analysis of the automotive market reveals the presence of four mega-trends that
are shaping the future of the automotive industry. The car of the future in fact will
be connected, autonomous, redefined and electrified.

1.1.1 Connected
Connected cars not only provide better experiences for drivers but also open new
ways for business to create value. Conventional vehicles will evolve into information-
enveloped automobiles that offer drivers and passengers a range of new experiences,
increasingly enhanced by artificial intelligence and intuitive interfaces that far sur-
pass today’s capabilities.
The key success factor for connectivity services is the clear value proposition the
offering has, either to an external customer or to an internal stakeholder. It seems
that this value is very often created only by combining data assets and capabilities
from various partners.

Figure 1.1: Percentage of consumers who feel that increased vehicle connectivity will
be beneficial

2
Connected car benefits are also perceived from European consumers, as it is possi-
ble to see in Figure 1.1. What is important to underline is that even the European
countries with the smaller percentages still feel that the increase of connectivity will
be beneficial.
Moreover, different types of analysis, just to name one, the McKinsey’s analysis,
have identified five levels of connectivity, each involving incremental degrees of func-
tionality that enrich the consumer experience, as well as a widening potential for
new revenue streams, cost savings, passenger safety and security, see Figure 1.3.
These levels reflect the potential for connectivity to stretch from today’s increas-
ingly common data links between individuals and the hardware of their vehicles to
future offerings of preference-based personalization and live dialogue, culminating
with cars functioning as virtual chauffeurs. The research suggests that by 2030, 45
percent of new vehicles will reach the third level of connectivity, representing a value
pool ranging from $450 billion to $750 billion. The surveys also indicate that 40
percent of today’s drivers would be willing to change vehicle brands for their next
purchase in return for greater connectivity.

Figure 1.2: By 2030, 45 percent of global new-car sales could be at level 3 or above in
connectivity.

Many manufacturers and suppliers already access a wealth of vehicle data to im-
prove or refine their cars and services, and possibilities abound for other players to
share information as new ecosystems form.
For these reasons, connectivity can be considered the technology enabler of all the
other mega-trends identified, especially with regard to the Big-Data that are col-

3
lected by connected cars, which will be discussed more in depth in the next chapter.

1.1.2 Autonomous
For investors, executives and enthusiasts alike, autonomous technology and self-
driving cars have long been some of the most interesting areas within the future-of-
mobility space. And even if, in 2019, progress in AV technology was not as fast as
expected, the underlying logic for autonomous driving, especially in cities, remains
intact. There is still the belief that electric, shared AVs, also called robo-taxis
or shuttles, could address mobility’s pain points in cities (such as road congestion,
crowded parking spaces, and pollution) while revolutionizing urban mobility, making
it more affordable, efficient, user friendly, environment friendly, and available to
everyone. If integrated seamlessly in the public-transportation system, it will be an
important enabler in reducing today’s share of private-car traffic.
Robo-taxi and shuttle mobility in fact, have the potential to disrupt our future
mobility behavior and to cannibalize many of the miles people travel each day. This
could fulfill daily mobility demands but also may signal the end of mass private-car
ownership—at least in high-income urban and suburban areas.
Customer adoption rates for robo-taxis and shuttles will vary by mobility use case:
customers will use robo-taxis and shuttles in different mobility use cases, most likely
with different frequencies.
The different adoption rates depend on convenience factors, such as finding a parking
space (or not) when going to the city center, and cost calculations, which depend
on the next best alternative for the respective journey. However, from the survey
carried out from Deloitte in 2019, emerged that almost 50 percent of consumers
agree that autonomous vehicles will not be safe, Figure 1.3. Therefore the way to
autonomous vehicles still seems long.

Figure 1.3: Percentage of consumers who agree that autonomous vehicles will not be
safe

4
1.1.3 Redefined
Furthermore, McKinsey’s analysis shows that global automotive revenues will nearly
double to EUR 5,500 billion in 2030 and will mainly originate from disruptive busi-
ness models such as mobility as a service (MaaS) or data-enabled services. At the
same time, profit pools will shift even more towards new technologies and services,
with more than 80 percent of the industry profit pool originating from the mega-
trends technologies and new business models. Therefore, the European automotive
industry will have to secure control points to maintain a profitable position in the
future and to participate in the changing revenue and profit pools.

Figure 1.4: While the challenges are significant, they in turn provide great opportunities
for players to conquer new markets and further reduce costs.

5
1.1.4 Electrified
Electrification—certainly gained momentum in 2019. This development was trig-
gered by two trends: tightening regulation—for example, in Europe—and rising
customer demand. In fact, as it is possible to see in Figure 1.5 consumers in Europe
are willing to switch from an ICE powertrain to an alternative one, preferring an
hybrid electric vehicle when considering their next car.

Figure 1.5: Consumer powertrain preferences for their next vehicle (2020)

From the same survey, it emerged that the reasons why consumers consider buying
hybrids or BEVs in Europe are in the first place the lower emissions and the lower
vehicle operating costs, as shown in Figure 1.6

6
 Figure 1.6: Reasons consumers consider hybrids or BEVs

There is no debating that the next five years will be a challenging transition pe-
riod for automakers and suppliers alike. Consumers, city dynamics, regulators, and
competitors will increase pressure on most OEMs to switch more quickly from ICE
vehicles to EVs, often with little consideration of EV economics.
Therefore, a clear road map of alternative powertrain is needed, including optimized
ICE/alternative fuels, electrified vehicles, hybrids, and fuel-cell vehicles towards the
2050 zero net impact emission target. Furthermore, a use-case based approach to
alternative powertrains could help identify the optimal powertrain for each mobility
use case with regard to local and total emission performance, mobility cost, cus-
tomer convenience, and regulatory requirements.
The challenge of making EVs profitable remains, but OEMs and their suppliers
are working hard to address it successfully. Advancements in battery technology,
economies of scale in EV production, native EV design, and cooperation among
OEMs can help bring down costs—which are currently still higher than for compa-
rable internal-combustion-engine (ICE) vehicles.
As starting point, since the business case is more attractive, OEMs are focusing on
large and medium-sized cars for the coming years. This is understandable from an
economic point of view but will not necessarily help OEMs meet CO2 targets at
scale, as the price point is still too high for many consumers.

7
Today, in fact, most OEMs do not make a profit from the sale of EVs. These
vehicles often cost $12,000 more to produce than comparable vehicles powered by
internal combustion engines (ICEs) in the small- to midsize car segment and the
small-utility-vehicle segment. What is more, carmakers often struggle to recoup
those costs through pricing alone. The result: apart from a few premium models,
OEMs stand to lose money on almost every EV sold, which is clearly unsustainable.
During the next five to seven years, as the industry transitions toward electrification
but struggles with profitability, automakers should more strongly consider partner-
ing and collaborating with competitors. These alliances will also be most beneficial
when they enable higher-volume procurement of the same battery cells and power
electronics to take advantage of scale that is otherwise elusive when going it alone.
In fact, some automakers have already announced a range of different global part-
nerships focused on reducing the cost of designing and producing EVs. According
to McKinsey’s analysis the impact of two OEMs codeveloping a dedicated EV plat-
form, which could lead to two to three times the volume spread across a similar
fixed-cost base-reducing costs by $1,500 to $2,000 per vehicle.
However, accelerating EV profitability will require some bold steps, including the
following:

• making tough choices around EV-platform design, including balancing lower

material cost with higher capital allocation and maximizing volume where
possible

• applying more ambitious cost-reduction approaches to EVs, including design

simplification, value-neutral decontenting, and aggressive purchasing strategies

• evaluating new potential partnerships with competitors to share R&D, tooling,

and production costs for new EV platforms

• considering more creative use of alternative EV-specific business models that

can boost margins

However, infrastructure needs to grow in line with growing EV demand. Fifty per-
cent of potential BEV buyers are concerned about limited range or access to charging
stations, as it is shown in Figure 1.7 where a recap of consumers considerations about
BEV/PHEV vehicles is reported.

8
 Figure 1.7: Consumers considerations about BEV/PHEV vehicles

Moreover, projections for Europe indicate that automakers would need to sell up to
2.2 million EV units in 2021 alone to meet their fleet CO2 targets. This is a steep
ramp-up of EV sales in less than two years and equivalent to global EV sales in
2018. This is a big task not only for the automotive industry, but also for adjacent
industries.

These four are the mega-trends that currently exist in the automotive sector and
since the entire thesis work is based on the study of Big-Data, the greatest enabler
in this direction is to be connected, so the Connectivity. Being connected is critical
to grouping and analyzing data, in fact nowadays it is an area in which OEMs will
invest a lot of resources and money, thus, this emerging new trend in connected car
data will be analysed in detail in the next chapter.

9
Chapter 2

The Importance of Big Data

The collection and usage of consumer data first entered public discourse en masse
in the early 2000s during the big tech boom as start-ups became multinational
powerhouses in record time. Much of this success was derived by providing a novel
service to consumers, such as social media outlets or intelligent web searching, in
return for unobstructed collection of users’ habits and interests data that was like
gold to advertisers, product designers, and more.
The automotive industry trailed the big tech data trends by more than a decade due
primarily to a lack of always-on connectivity, but also due to the slower pace of tech
innovation in the industry. Now that in-car connectivity is nearly ubiquitous, and
the industry has pivoted to a more agile software-driven structure, automakers are
rapidly exploring new opportunities to capitalize on the potential value of the data
collected. In many cases, this takes the form of providing new or enhanced services
to the customer, while in some cases, the data may be used for the greater good,
such as improving infrastructure and safety. While there are many potential uses for
in-car data, automakers must balance the potential value with consumers’ demand
for privacy, which is becoming increasing mandated by various regulations. With
careful consideration of their data usage, automakers can improve their product and
service offerings, while respecting privacy rights, and still enhance their bottom line.
What is clear to everyone today is that a clear value-add strategy is required for an
automaker’s data program to be successful. Connected car data was once touted
as creating instant profit, yet the reality is quite different. Data by itself generates
very little revenue, yet creating true insights and value-add services from the data
will put the automaker on the path to success. The implementation of this kind
of program may require organizational changes and significant resources to get the
program to the market.
The main questions that the management of this data have led to OEMs are multiple
and that still today have difficulty finding a real answer are the following:

• How to define success for my organization?

• Are laws threatening opportunities?

• How do we catch up to the competition?

10
2.1 How to define success for my organization?
The success of a connected car data program comes in many shapes and sizes and
will be determined at least in part by the automaker’s personality, challenges, and
organizational goals. Regional differences also play a large role in defining success
due to various regulatory issues, differences in consumer willingness to pay, and dif-
ferent market maturity levels. The best place to start is to ask what your customers
need and want, and then determine how those do or do not align to your regional
and global goals. Some additional goals may not directly involve your customers and
so a second round of goal setting should look at internal needs. However, the au-
tomaker should be careful to understand if any of the internal goals could also affect
the customer and whether that is in the customer’s best interest. Automakers typi-
cally define several types of goals, including both internal goals and customer-facing
goals. Regardless of what the actual goals are, success cannot be determined with-
out measurement of KPIs. A regular PDCA (Plan, Do, Check, Act) cycle should
be incorporated into any new data program. Some of the most common success
definitions are shown below:

1. Champions of the Greater Good: Many automotive data sets can be

leveraged in a way that has an immediate and direct impact on broader social
issues. The simplest examples are related to safety, where road ice detection,
emergency braking, etc., can be relayed to other cars in the area or (in the case
of ice detection) to the highway authorities. Another example is using a car’s
driving behavior data to detect road designs that are particularly dangerous.
These types of data are typically given away to public authorities as a form of
corporate social responsibility. While announcing the activities to the public
is encouraged, it should not be used as an advertisement, since this will cross
a moral boundary for many customers.

2. Facilitator of 3rd Party Services: Perhaps the most visible use of con-
nected car data is the enablement of services such as fleet management, usage-
based insurance, and listenership tracking. While the 3rd parties will make
a clear case for a synergistic relationship, very few of these partnerships yet
develop significant consumer uptake or revenue for the automaker. An over-
abundance of 3rd party offers may also confuse the consumer or taint their
perception of the brand.

3. Warranty cost reduction: Many automakers begin their automotive data

journey by analyzing trend data from specific components, such as batteries
or fuel injectors, to predict failures. On paper, this seems like a straightfor-
ward concept, yet in practice can require teams of data scientists to turn the
mountains of data into reliable insights. By detecting design flaws prior to
in-market failures, automakers can dramatically reduce their warranty costs
while improving consumer satisfaction, and potentially avoid recalls. The pre-
dictive maintenance is one of the much discussed topics today by all the most
important OEMs in the world, even Lamborghini is a topic on which it is
reasoning and on which in the future it would like to focus.

11
2.2 Are laws threatening opportunities?
At first glance the answer to this question is YES, however, the various types of
legislation should not be viewed as a homogeneous group with similar implications
for the automaker. Data privacy legislation for instance should be viewed as an in-
evitable social initiative that supports human and competitive market rights. Other
forms of data legislation such as Right to Repair laws or the EU’s upcoming Data
Act, are not quite so clear cut, and may lead to a paradigm shift in automotive and
other connected device industries.

2.2.1 Data Privacy Legislation

By their nature, data privacy movements will infringe on some of the existing op-
portunities that automotive data can provide. The development of more restrictive
privacy legislation will continue globally until consumers and governments feel that
their natural right to data privacy is wellrepresented in law. However, this does not
mean that automotive data programs are doomed to fail, nor that they won’t provide
real value to consumers, automakers, and 3rd parties. Using Europe’s GDPR as an
example, the legislation is quite restrictive yet creates a framework for the respon-
sible handling of personal data, and only with explicit consent from the customer.
This ensures that car owners aren’t taken advantage of and provides automakers
and 3rd parties with a supportive framework for the creation of value through data
analysis and trading. Privacy legislation such as this will certainly add overhead to
the development of any form of connected vehicle service yet will result in an or-
ganization that works to support their customer’s rights, ultimately improving the
customer experience. The services that are developed under these privacy frame-
works will still create value, but in a manner that is completely transparent and
consensual.

2.2.2 Right to Repair

This divisive legislation has seen significant activity in the last 10 years, with the
European Commission and individual states in the USA mulling over various ideas
about how to democratize vehicle data so that aftermarket repair shops and 3rd
parties can gain access without going through the OEM. While both sides of the
argument have merit, it is clear that this is a threat to the automakers’ traditional
perception of being data owners. The advent of 3rd party data marketplaces (and
automakers’ proactive engagements with them) may have delayed the issuance of
new legislation by several years, however new regulations come into effect in 2023,
requiring a standardized data interface.

2.2.3 European Commission’s Data Act

This legislation, which is still in the proposal phase, could result in the broad reclas-
sification of non-personal connected device data as being a public asset. This could,
for instance, allow any third party, including competitors, to analyze all non-personal

12
data that the automaker receives from the vehicle. This act could indeed usher in a
new wave of economic development based in data science and data-for-good initia-
tives yet would strip automakers of many current opportunities and eliminate any
advantage currently held.

2.3 How to catch up with the competition?

While many automakers have mature vehicle data programs, some have yet to cap-
italize on the opportunities presented by connected data. The maturity of an au-
tomaker’s data programs are typically proportional to the share of vehicles that
they sell with embedded modems. Just a few years ago, many high-volume brands
didn’t offer embedded modems on any of their models, instead opting for smartphone
tethering and mirroring solutions. This is partially because these automakers saw
embedded connectivity simply as a premium feature that would need to be reserved
for their premium lineup where customers would be more likely to purchase the fea-
ture. While the practice of ‘de-contenting’ is common among an automaker’s brands,
it also artificially constrained the creation of value by ensuring that the connected
vehicle fleet was limited to low-volume premium vehicles. Furthermore, because few
value-added services were being offered to the customers, subscription rates were
typically very low, further compounding the problem. The final twist is that many
of these automakers tie the connected services budget to the revenue generated by
the subscriptions and add-on services. This organizational design guarantees that
the chicken-and-egg problem will never be resolved through natural growth. Now
that the problem faced by many automakers has been examined, you will see how
an immature or non-existent data program can become competitive in the industry.
1. The first step is to increase the penetration of embedded connectiv-
ity in your fleet. This is an investment in the company’s future. While this
will add cost to the BOM, it will improve customer satisfaction and enable the
additional use cases that provide both monetary and non-monetary value to
the automaker.
2. Provide the customer with a long (3+ years) complimentary con-
nected service period combined with well-designed, simple features.
The simple, well-designed service will ensure that customers renew their sub-
scriptions. If the internal value created by the platform is large.
3. Don’t try to run before you walk. Attempting to stand-up too many
customer-facing services or data use cases at once will likely lead to none of
the services receiving the design attention that they desire, will require a larger
initial investment, and may lead to a disjointed data organization. Develop
internal data use cases and 3rd party partnerships gradually, ensuring goal
achievement of existing initiatives prior to moving on.
4. Ensure that the connected car data organization starts with a ro-
bust governance model with support from top executives and the
authority to pull internal ‘levers’ to meet KPI goals. For instance, the
team should be empowered to explore innovative data use cases to optimize
value creation.

13
These concepts are very much based on the four mega-trends mentioned and an-
alyzed by previous chapter, in particular the trend related to Redefined the au-
tomotive revenues with the creation of Business models completely different from
before, with nowadays the need to have models such as mobility as service (MaaS)
or data-enabled services.

2.4 Who’s who in the automotive data industry

The automotive data landscape is divided into the suppliers of data (the automakers)
and the suppliers of data services (the data platforms). SBD’s analysis reveals that
while there are many data platform providers, the majority seem to have slowed
their development and client onboarding, leaving just a handful of companies to
pursue the large partnerships with OEMs. For some of these companies on the left
side of the Data Platforms chart, automotive data processing and brokering is a
small portion of their business and therefore they may not intend to pursue more
than a basic level of capability and business engagement. Meanwhile the focused
companies, such as Wejo and Otonomo are competitively pursuing new business.
The Historical Trend arrows show the general movement that SBD has witnessed
over the last few years. The automakers on the other hand, are all moving in the
same direction, just at different speeds and with different strategies. We expect the
cluster of automakers in the middle of the chart to continue moving to the top right
to remain competitive.

Figure 2.1: Data Platforms

14
 Figure 2.2: Automakers

OEMs can potentially realize enormous benefits from knowing which parts of a vehi-
cle are likely to fail and when. Real-time data sent from vehicle sensors can identify
problems early, and predictive analytics can allow companies to get out in front of
potential warranty and recall issues. This kind of data can also help OEMs and
dealers optimize their parts inventory and technician resourcing strategies. Further,
a deeper understanding of how customers use their vehicles can help OEMs design
better, more customized customer experiences to improve brand affinity and loy-
alty. In terms of external revenue opportunities, OEMs and other industry players
are exploring a wide variety of data-based products and service offerings, including
user-based insurance, mobile commerce, mobility-as-a-service (MaaS), behavioral
and geo-based advertising, infotainment, and personal health monitoring. However,
many OEMs are already showing their collective vulnerability to new and existing
entrants looking for ways to go around them. In the case of user-based insurance, for
instance, the addition of a simple plug-in allows insurance companies to gain access
to vehicle usage data, thus circumventing the need to interface with OEMs. Vehicle
manufacturers may also face an uphill struggle with significant consumer concern
over the collection of bio-metric data. For example, recent survey data suggests that
63 percent of US consumers are at least somewhat concerned about bio-metric data
being captured and shared with external parties. The complexity and dynamism
that characterize the emerging connected vehicle industry has made it difficult to
make decisions regarding where to play and how to win; in fact as it is illustrated
in the figures 2.3 and 2.4 , it is possible to see what are, at the moment, the main
drivers and barriers that are affecting the connected car data.

15
Figure 2.3: Drivers affecting connected car data

Figure 2.4: Barriers affecting connected car data

16
Obviously, looking at the two figures, we can well understand that the drivers that
influence today and that will influence in the future will be more an exponential
growth of connected cars, also there will be in the near future a little change in the
mindset of the people, in fact 64% of consumers surveyed answered they are likely
to share their data for automotive use cases. However, just when it comes to the
management and sharing of customer data, the first barriers are born; in particular,
OEMs must clearly explain to their customers how and why their data is being
collected and, moreover, they must provide customers with complete control over
what data is shared and to whom, otherwise, if there is not this transparency and
clarity on the part of OEMs, what is at risk is that customers will have less and less
confidence in the management of their personal data. Beyond consumer demand for
data privacy, there will also be a huge increasing number of data privacy regulations
across the world and for that reason the OEMs will have to comply with the strictest
data privacy regulation and for sure, Data Lake processing must be local in some
regions: i.e. Data Act in UE, China etc.. Putting these main drivers and barriers
on the scale and due to the fact that the compounding data volumes will quickly
overwhelm an unprepared organization, we have to ensure the presence of an ideal
structure, budget and strategy are in place, otherwise the management of this data
will become a real hell.

2.5 Data monetization opportunity

Connected services and data monetization have long been seen as a big opportunity
for automotive companies facing an uncertain future. However, the path to success
may require new thinking in terms of where to play and how to win. For decades,
automotive original equipment manufacturers (OEMs) have been able to find suc-
cess in the relative comfort of building assets for personal consumption. But outside
the industry, companies in areas such as financial services, biotechnology, and social
media were hard at work finding new ways to generate revenue and grow shareholder
value. Many biotechnology companies, for instance, found their niche as “technol-
ogy creators,” while social media players and other “network orchestrators” created
asset-free revenue models based solely on bringing supply and demand together in
a market exchange environment.
Although OEMs have introduced countless innovations over the years, they have
never really been credited as “technology builders” by the investment community.
In fact, their valuations still hover around those of the lowest “asset builder” quad-
rant. So how can traditional automotive companies start moving up the revenue
multiplier ladder? The answer may be found in new business models that promise
to leverage the increasing amount of data being generated, captured, and shared by
the vehicle itself. Vehicles are now able to capture and share many types of data,
including geolocation, vehicle performance, driver behavior, and biometrics data.
Though GPS functionality has supported navigation systems for years, smarter ap-
plications of the data are adding significant value in the form of real-time traffic
updates and road safety alerts. Uses for vehicle health and operational functional-
ity data are also spreading as vehicle manufacturers continue to develop app-based
tools to monitor key maintenance statistics. And while the use of advanced bio-
metric data is still in its infancy, new sensors in the cockpit can allow vehicles to

17
monitor key attributes of their occupants, including stress levels, heart rhythms,
alcohol consumption, and fatigue. However, monetizing this tremendous increase in
operational and behavioral data is easier said than done—and OEMs have largely
been lagging behind market disruptors entering this space. There is also good reason
to wonder whether a critical mass of consumers see the increase in vehicle connectiv-
ity as a good thing. Recent study results suggest that while 79 percent of consumers
in China believe increased connectivity will be beneficial, only 35 percent of German
consumers feel the same.
One of the first decisions for companies aiming to monetize vehicle data is where to
play in the connected vehicle value chain. Potential roles exist for companies to act
as:

• Generators: making end products capable of capturing data;

• Transmitters: safely delivering the data to a central repository;

• Manipulators: aggregating data from different sources into a usable format;

• Developers: designing end-user offerings that leverage the data;

• Providers: marketing the service offerings to both Business-to-Consumer

(B2C) and Business-to-Business (B2B) audiences.

Not every company is well-placed to succeed in each part of this value chain. For
new entrants in particular, it can be difficult to create value further down the chain
without access to the data being generated further upstream. Here lies one of the
central issues that is currently preventing the vehicle data monetization ecosystem
from developing to its full potential. Many OEMs have the data, but because they
want to control every point in the value chain—even though they are not generally
well positioned to do so—they are reluctant to make the data available to anyone
else.

2.6 What is Data Monetization and why are au-

tomakers trying to monetize vehicle data?
Data monetization is the practice of using data to derive revenue. This business
model can take many forms, from the immediate return of selling data directly to 3rd
parties for a fee, to the long-term practice of providing value-add services for free in
hopes of increasing repeat purchases. OEMs currently have the greatest opportunity
to capitalize on data monetization use cases since they have direct access to the full
set of vehicle data and access to the customer. Where OEMs aren’t able to provide a
service, they can choose to sell the data directly to 3rd parties or offer services to the
customer in partnership with the 3rd parties. Data privacy legislation is beginning
to make it more difficult for OEMs to proactively offer services to their customers
without fear of legal action. Some 3rd party hardware vendors are attempting to
compete with the OEMs by offering services powered by OBD dongles. The picture
2.8 shows briefly how is made the ecosystem in which the data flows everyday.

18
 Figure 2.5: Data ecosystem overview

As the automotive industry evolves, automakers are becoming increasingly focused

on monetizing vehicle data. Automakers are investing in data monetization for five
main reasons:

1. Increased desire for autonomous vehicles, shared mobility, fleet management

and usage-based insurance opens new, data reliant business opportunities for
automakers.

2. Increasing RD costs for Evs, autonomous, connected platforms and manufac-

turing are eating into automaker’s profits.

3. Increasing complexity with the automotive industry is impacting decision mak-

ers trying to keep pace with consumer desires.

4. Increased vehicle electronics and component value are negatively impacting

manufacturers servicing costs.

5. Increasing competition from industry disruptors means manufacturers must

develop higher quality vehicles and services.

19
 Figure 2.6: Why automakers are trying to monetize vehicle data

Typically the Big Data monetization environment is divided in two main streams:
the Internal Data and the External Data.
The Internal Data monetization refers to the use of data for internal benefits, notably
process efficiencies made possible through timely trigger and/or the reduction of
assumption through rel-time visibility to the connected fleet. This requires clear
business requirements, careful handling of personally identifiable data and a means
to access the data.

Figure 2.7: Internal Data monetization

Whereas, the External Data monetization refers to selling or trading connected car
data to third parties outside of the originating organization’s boundaries. The most

20
typical use cases today are established by providing third parties with a reduction in
costs, as they can trade dongles and asset management for a contract or subscription
to TCU data.

Figure 2.8: External Data monetization

However, the real question is: ”Why is there a need for a vehicle data monetization
ecosystem?”. Traditionally, OEMs focused on vehicle design, manufacturing and
retailing; in the last ten years, OEMs have struggles to provide customers with the
consumer electronics-type of technology in vehicles, an area outside the traditional
core competencies; instead now that vehicles generate potentially valuable data (the
raw resource), there are opportunity to manipulate, consume and selle this data.
OEMs have neither the resources or experience to turn this into a refined product for
consumption and and for most of OEMs, this should be delegated. Essentially, data
monetization is an activity in a different “vertical”, or industry, than automotive.
Just as an IT company would struggle to manage the vast parts supplier network
and supply chain management of an OEM at a global level, OEMs struggle with the
data ecosystem. Every OEM has developed its connected car solution and associated
data environment independently. At the global level, this has dozens of incompatible
data sets that third parties want to access, each changing and evolving over time.
The end consumers of this data do not want the burden of integrating independently
with each OEM. Therefore, a demand has arisen for platforms that can act as the
broker between OEMs and developers, providing easy monetization opportunities
for OEMs and scaled data access for developers. In fact, nowadays the vehicle data
monetization ecosystem has this shape, as we can see in the figure 2.9.

21
 Figure 2.9: How the current vehicle data monetization ecosystem looks like

The OEM can choose to expose data externally in the manner that bets fits their
goals and desired business model. There are two general ways OEMs expose data:

1. As a data provider to data marketplace service providers who then manage

the integration with various 3rd parties;

2. As a data provider to individual 3rd parties using an in-house API or integra-

tion model.

While marketplaces may offer better scale, 1:1 relationships with developers can
create unique, value-add services curated based on strengths of both the OEM and
the 3rd party. Data Marketplace generally treat their platform user as either data
providers or data consumers. Data providers to marketplaces provide vehicle data
and are generally automotive OEMs, or in some cases telematics service providers
and fleet providers. Data consumers are the parties that use the vehicle data which
is generally accessed through a common API.
In basic terms, data provides the opportunity for triggers (when to do something)
and insights (visibility to a situation or condition). Therefore, data helps to fuel
business processes and these processes must be well defined and implemented for
the data to provide its value. This means that there are a number of steps required,
calling for varying areas of expertise and resources. As a result, different experts
are required to perform specific steps, creating a value chain. Some parties are

22
well-suited to handle parts of this value chain. However, others are foreign and
complicated, requiring partnerships, thus creating an ecosystem. In recent times,
a new area of the connected car ecosystem has opened up, comprising startups
or new departments of entrenched industry players who provide “data services”.
These companies provide a wide array of services such as integration into data
marketplaces, such as Otonomo, wejo or Continental’s blockchain-enabled platform;
white-label data server products such as IBM or Ericsson and API development
tools such as HIGH MOBILITY or Smartcar. OEMs who are pursuing a data
monetization strategy have taken to partnering with multiple companies as it is still
unclear who or through which model future success will come.

2.7 What does the future hold for automotive

data?
The future of the automotive data industry will be shaped by several external fac-
tors over the next few years. The most notable is the regulatory landscape which
will create many hurdles for automakers, yet these changes should be seen as a nec-
essary step to ensure a level playing field globally and to provide fair treatment to
consumers. In the more distant future, it’s likely that technology plays a greater
role in shaping the service model, with advanced 5G (and 6G) networks and edge
processing enabling new use cases that had once not been feasible.
More precisely, we can summarize the future of automotive data in the following
main pillars, making a bit of a prediction on what the future scenarios may be:

• Regulations:The European Data Act will be coming into effect in 2025 cre-
ating a paradigm shift in the ownership of data. In addition, the EU’s Data
Spaces will begin to spur commercial growth in the data analytics industry.
The USA will also likely enact federal legislation that provides a similar form
of data protection and portability. Automotive data analytics and services
now involve governments, researchers, and service providers working in many
other sectors including public health safety, power generation distribution,
road infrastructure planning, and predictive mobility services.

• Consumer Needs: Consumers will begin to demand portable digital profiles

that follow them from their smartphone to any car, laptop, or other inter-
active device. Certain services such as traffic information and map updates
will be mandatory. Usage-based-insurance may begin to gain in popularity
if it is provided seamlessly. Consumers will have become familiar with the
concept of managing their data rights and they will ensure their rights are not
infringed. They will expect seamless profile portability between products and
platforms. Micro transactions for connected services are now more common
than subscriptions.

• Technology: New platforms will feature more centralized electrical archi-

tectures enabling greater interaction between the vehicle’s systems and the
back-end connected car platform. OTA updates are ubiquitous for major com-
ponents. Higher power computing is beginning to shift processing from the

23
 cloud to the edge. High power computing is ubiquitous, enabling onboard pro-
cessing of complex data including video and LiDAR. Insights from the data
can be processed onboard and transmitted to the OEM’s backend platform for
AD/ADAS model improvement. Commonplace 5G enables backend systems
to react to live data in real time.

• Automaker Business: Most major automakers have some form of connected

vehicle platform in place, yet only a few have developed sustainable internal
and 3rd party use cases. Those that are still developing their connected vehi-
cle business will likely emulate services already in place by competitors. Some
OEMs will choose to set up not-for-profit data programs, rejecting moneti-
zation completely. 3rd party services such as aftersales service, usage-based-
insurance, pay-as-you-drive, and mobility subscriptions will be fiercely com-
petitive with the OEMs’ own services, forcing the OEM to leverage brand
loyalty, perceptions of quality, and convenience, to win the business.

• Data Platform Business: Data collection, aggregation, and brokering are

becoming more difficult as regulations such as the EU’s Data Act drive down
the value of vehicle data, and analysis paralysis plagues clients. The busi-
ness is now focusing on enabling 3rd party services, deriving unique insights,
and consulting services. Smart city and mobility services are finally becom-
ing a reality with Data Platform Providers serving as a critical link between
automakers, various mobility services, and government bodies. Data collec-
tion and aggregation is fully commoditized, yet deep insights leveraging edge
processing and ML, are prized.

In conclusion, in order to give a more precise and numerical outcome regarding

the future perspective of the automotive data, see 2.10, we can finish saying that
between 2020 and 2025 automotive data volumes are expected to increase 1000-fold.
Even when considering the effect of falling mobile data costs, this could lead to
automakers facing data bills as high as $300/month.

Figure 2.10: Data volumes and costs could skyrocket by 2025

Clearly this is not a sustainable business model, meaning that to take advantage
of the large quantities of data generated by these vehicles, automakers will need

24
to consider new strategies such as the generation of data insights on the vehicle
instead of in the cloud. This could take the form of new high-power edge processing
architectures delaying transmission of certain data until the vehicle is able to connect
to a WiFi access point. The latter strategy would work well for transmitting sensor
and vehicle data to train machine learning algorithms, or perhaps uploading road
condition data to local road authorities.

25
Chapter 3

Lamborghini Projects on Big Data

All the mega-trends and mutations that are affecting the automotive world has also
led Lamborghini to be forced to follow these new changes in order to always keep
up with the new needs of the market without, however, descending from its roots
and its DNA which has always been distinctive among all the OEMs in the world.
Therefore, Lamborghini has also begun to carry out several Proof of Concepts (PoCs)
related to the analysis of Big Data of connected cars that it has already placed on
the world market, which have the main goal to understand how derive value from
these data and, in particular, what kind of data, among all those available, can
be more useful both from a revenue and strategic point of view for the company.
The study of data, as already seen in the previous chapter, leads to a new business
model approach, profit pools shift even more towards new technologies and services;
in fact, for this reason Lamborghini has undertaken the study of the connected data
in order to understand what are the main driving habits of its customers to offer
them services more suited to their type of driving.
This type of approach finds two main benefits, first of all improves by far the cus-
tomer experience and, secondly, the study on the Big Data can bring a great advan-
tage also to the company, because it has the power to focus more on what customers
really want rather than on less flashy things. Therefore, the different projects on
which Lamborghini has been focused in the last year will be discussed and on each
project there will be a focus on the quality of the KPIs that have been reached.
All the PoCs made are not based on the same inputs,in particular all the PoCs are
based on the same data stored on SDP which is the central point of the Lamborghini
Huracan connected vehicles and it manages all the data sent by the cars, IT systems
and customer touch-points, but we must make a distinction among the four PoCs
considered; in particular three PoCs are focused on a static data-set in which all
the data that had been stored from January 1st to November 29th, 2021 by 1457
connected Huracan; instead only one PoC is based on the same SDP data but on an
extraction, in particular all analyses are based on the last five months’ data stored
from 457 different vehicles. Furthermore, both considering 1457 vehicles and the
extraction of only 457 vehicles, a static data-set was always used, this choice was
made for simplicity and in particular, for the purposes of the PoCs, it was not im-
portant what type of data-set is used, whether static or dynamic, but the main goal
is to understand how to extract value from these data and define new KPIs from
Data for Connectivity.

26
3.1 First PoC launched by Lamborghini
The first work carried out by Lamborghini lasted about four months, from October
2021 to January 2022 and as previously mentioned the entire work focused on the
analysis of static data saved on a historical database from connected fleet from
January 1st to November 29th, 2021 by 1457 connected Huracan. In this first
study, the main goal was, first of all, identify user types using the data collected
by connected cars; second, develop an algorithmic solution that allows the car to
be associated with a user type. Obviously not all the information was taken into
account for the creation of the clusters but only a few, in particular the information
used for the creation of the clusters are six:

1. Maximum speed;

2. Minimum speed;

3. Frequency of use;

4. Duration of car use;

5. Total distance traveled;

6. Number of cars taken into account.

Additionally, the days of use were calculated as a fraction of the days of use at
speeds greater than zero on the days elapsed between the first observation of the
dataset and the last, November 29th. The maximum speeds have been calculated as
the average of the maximum daily speeds of each car. The kilometres now represent
the average number of kilometres travelled per day. The driving hours per day refer
exclusively to the series that the algorithm has considered useful for the analysis,
thus excluding small movements not very relevant.
So on the basis of these six parameters, it was made a cluster analysis. Cluster
analysis involves the autonomous discovery of clusters in the data based on patterns.
This type of statistical technique allows to assign vehicles to a group of similar cars
on the basis of common characteristics without the need of preset categories. At
the end of the analysis, five clusters have been mathematically found that describe
the different types of behavior that can occur when a person is driving. So, the five
clusters, as shown in the table 3.1, identify the possible ways in which a driver can
use his car in everyday life, namely:

− semi-daily use to go to work every day;

− use in the city and on weekends;

− periodic use on the track ;

− extra-urban use on weekends;

− extra-urban use on weekend.

27
By studying the data coming from a given car it is possible to immediately under-
stand in what kind of cluster it falls into.

Figure 3.1: Lamborghini cluster customization

In addition, a further analysis was also made to verify the cluster customization
made previously. In the figure below 3.2, it is possible to see a certain number of
machines near the Misano circuit and on the basis of the parameters recorded by
each vehicle it is possible to group that particular car directly to one of the five
clusters seen above.

Figure 3.2: Visualization of the clusters

In this case, the realization of the different clusters, done previously, by means of
special algorithms based on scientific clustering techniques, brought a very good
result, because intuitively the colors that appear on the circuit are reliable, given
the fact that all the dots that are in the circuit belong to the same cluster and those

28
that are in the surroundings of the circuit belong to diversified clusters which makes
absolute sense.
Finally, the study ended with a further analysis, that is, using the data referring to
the total distance traveled and duration of car usage, a forecast was made on the
average distance traveled by each vehicle between two journeys. What emerged was
that on average the distance traveled between two journeys is the same for all five
types of clusters and approximated to a value of 4 hours (3.3).

Figure 3.3: Big Data as enabler for the electrification

Following the line of the previous analysis, it was also estimated how many kilome-
ters on average are traveled each trip, and it emerged that on average never exceed
78.2 kilometers. These last two analyzes, in view of a future in which the electri-
fication of cars will be increasingly present, among which Lamborghini will also be
the protagonist, these data could be taken as a starting point in deciding what kind
of batteries a Lamborghini of the future will wear. These last considerations could
be of great value for Lamborghini, especially in view of the great changes that the
automotive world will see in the coming years in which it is planned, as already
stated by the European Commission, to stop cars with internal combustion engines
by 2035, switching exclusively to the electric mode.

3.2 Second PoC launched by Lamborghini

The second project on which Lamborghini has focused, lasted about five months,
from October 2021 until the first week of April 2022, and the whole work was
always focused on the same data-set as the 1457 connected Huracan, like the PoC
seen before, but in this case the results that have emerged are very different from
those seen previously.
Regarding this PoC, I have been an active part in the decision-making process so
in this chapter we will only see the main data that have been examined and the
general outputs that have been obtained; I will talk about this PoC in more detail
in the next chapter where each analysis and parameter obtained will be examined
in great detail.
In comparison with the previous study, in this case the input data were not taken
individually, but were grouped as follows:

29
 1. Vehicle Dynamics;

2. Driving Settings;

3. Powertrain;

4. Geo-localization;

5. Weather Data.
After careful filtering and cleaning of the input data, several outputs were obtained,
which opened the door to interesting use cases and important KPIs that Lamborghini
can use in the decision-making phase in the strategies of the future. The outputs
that have been obtained are as follows:
• Geo Spatial distribution of vehicles (World region, country, cities);

• Usage analysis on the entire fleet, single vehicle and single journey;

• Driving mode clustering and characterization;

• Correlation between driving habits and weather conditions;

• Race track visits;

• Service Shop visits;

• Interactive Dashboard for data reporting.

As already mentioned above, in this chapter, regarding the following PoC, I limit
myself only to saying what were the parameters taken into account and what kind
of outputs were obtained from these, all the further details and insights will be given
in the next chapter, in particular, given that all the use cases addressed have been
put on an interactive Dashboard and, for that reason, the visualization of the uses
cases will make the idea of the work done even better.

3.3 Third PoC launched by Lamborghini

The third study carried out by Lamborghini lasted just under three months, from
November 2021 to January 2022, and here the data-set taken as reference is different
to the usual one seen so far, but instead is based on 457 connected Huracans and
focusing on only five months of the data that has been recorded, and indeed the final
outcome showed that the results emerged were very different from both analyzes seen
previously. In this case, an interactive dashboard that updates itself real-time has
been implemented.
Three different Use Cases were addressed in this PoC:
• Statistical Data Analysis: analyze the MQTT data sampled from the ve-
hicles in order to discover useful statistical data.

• Digital Marketing: analyze the MQTT data in order to propose new func-
tionalities or to invite the Customer to events.

30
 • Connected Services Usage: analyze the usage of the connected services in
order to invest in the development of innovative connected services.

Starting from the first use case, the main goal was to identify the most used driving
mode for each vehicle model and model year; in fact the parameters that have
been studied are data such as the kilometers traveled in each geographical area, the
driving mode used for each area and the kilometers traveled for each model and
model year in each area (3.4).

Figure 3.4: Statistical data analysis on Dashboard

After that, the second use case was related to the digital marketing goal, here the
focus was to analyze the driving mode combining with the vehicle configurations in
order to propose the telemetry functionality. As shown in the figure 3.5; for this
specific use case, the Dashboard shows the Lamborghini events where are geograph-
ically located and below are returned the whole list of cars that could be interested
in participating in these types of events. The list is created based on some criteria,
as the distance, so the single car will be put in the list only if it is in the vicin-
ity of the event otherwise it would be useless; the second criteria is based on the
driver’s driving style, maybe those who drive more cautiously will be interested in
different types of events than the type of customer who likes to drive the car more
aggressively.

31
 Figure 3.5: Dashboard for digital marketing

Finally, the last use cases was also extracted value from the connected services usage
(3.6). In this case, the dashboard shows different insights about remote services,
pairing status, and on/off privacy status. This last use cases were the ones of
greatest interest and usefulness for Lamborghini because, in real-time and in an
absolutely precise way, it is possible to see the status of pairing and privacy, making
everything super automatic and easy to manage.

Figure 3.6: Connected services usage Dashboard

In conclusion, the request for this PoC was born because the current tracking modes
are not yet collected on an ad-hoc platform/dahsboard by Lamborghini, but the out-
puts obtained showed excellent results in this direction having everything automated
on a single dashboard; therefore, at the same time it has offered new ideas, such
as the use case focused on digital marketing for events organized by Lamborghini
around the world. In general, this PoC has brought many interesting ideas to focus
on that will be evaluated very soon by Lamborghini.

32
3.4 Fourth PoC launched by Lamborghini
Finally, the latest study carried out by Lamborghini is a work that began in July
2021 and is still ongoing. This last PoC is structured in a completely different way
from those previously seen, as it is based on a broader and at the same time more
complex vision, all focused on the Predictive Maintenance use case. As can be seen
from the architecture, figure 3.7, the information are extrapolated not only from the
connected Huracans but also from other types of sources such as the data recorded
by the workshops and dealers, but why? Because the main object of this PoC is to
process the data from the connected cars and predict possible faults/failures or need
for servicing; but as can be guessed, to implement such a use case, it is impossible
to rely only on the data stored by the car, but there is the need also of external data
that can help the cause, such as the reports from the dealers.

Figure 3.7: Data Lake architecture

The benefits that could brought this PoC is the increasing of vehicle reliability and
reduction of the warranty costs; it would also allow dealers to be proactive in con-
tacting customers in case of a service/breakdown need. Currently there are several
problems in the implementation of the PoC, in particular the pain point is the dif-
ficulty to assess vehicle reliability and identify ways to reduce warranty costs.
Finally, in this type of analysis, through predictive analysis and advanced modeling
algorithm the need is to extract value from a huge amount of data, not only those
data coming from the connected Huracans, so the work behind it is much more
articulated and more complex than the PoCs seen previously, in fact it is also partly
for this reason that this project is still ongoing.

33
Chapter 4

Data Analysis and Dashboard Im-

plementation

As already mentioned in the previous chapter, in the last year Lamborghini has been
focused on different projects regarding Big Data, to understand how to really get
the most value from the data of the connected cars put on the market, in order to
improve the customer experience as much as possible and reach at the same time
the greatest economic gain.
As already anticipated, in this chapter I will directly investigate more in details a
PoC already mentioned in the previous chapter, in which I was an active part in
some decisions that have been taken.
Obviously, the PoC in question was always based on the static data-set referring
to all data taken into account from 1st January until 29th November 2021 of the
1457 connected Huracans; but now I will go, step by step, into the main phases in
which the data were analyzed and how they were correlated between them in order
to obtain valuable data.

4.1 Data Preparation

The starting ecosystem from which it all began is the one shown in the figure 4.1.

Figure 4.1: Ecosystem of the Architecture

34
As can be seen, the data that have been taken into consideration for the analysis of
possible use cases are various. The data relating to vehicle dynamics and powertrain
dynamics, such as acceleration, maximum speed, average speed, etc., were taken
directly from the static Lamborghini data-set; whereas, the data referring to the
meteorological conditions are not part of the reference data-set, but it was decided
to insert it later because it could be useful to extract other relevant use cases.
So, the first thing that was done is to work on data that make sense and that are as
correct as possible, so the first phase, as well as for all the other PoCs seen in the
previous chapter, was the preparation of the data and this had several steps, which
are those of the figure 4.2.

Figure 4.2: Data Preparation

The first step was to take all the raw data that were part of the starting data-set
and do a first VIN filtering. As previously mentioned, the total connected vehicles
taken into consideration are 1457 of which 46 were test machines, therefore in order
for the analysis to be as accurate as possible, the test machines have not been taken
into consideration because they are machines that are used for different purposes
and therefore fail to represent a typical use.
So once we only had the data from connected machines to actually work on, the
second step was to do a cleanup of this data. This cleaning was fundamental at the
end of the study, since in the starting data-set there were some measurement errors,
for example sometimes maximum speeds of 500 km/h were reported, or sometimes
there were gaps in measurement leaving the box empty; therefore this second phase
was very important because it made the data-set more coherent and more precise
in the measurements that were contained within it.
After that, the third phase was a little more articulated but at the same time very

35
useful. On the basis of the data obtained from the second phase just seen, macro
clusters have been created so that a division of the journeys can be created for a
better understanding of them. The clusters were created taking into account the
data that refer to the dynamics of the vehicle and the powetrian, such as speed and
acceleration, in order to differentiate the typologies of journeys.
In addition, as a fourth step, was made a GPS analysis in such a way as to further
clean up the data-set by considering only those countries that are really connected;
this type of preparation, as will be seen later, was very useful for different use cases.
Finally, the last step was relating to a weather analysis, in particular data were taken
regarding the weather conditions not of all cities around the world but only of three
cities: Milan, Miami and Los Angeles. These data were very useful for correlating
the use of the car on the basis of the weather conditions, in order to create new and
interesting use cases to be taken into consideration.
After this first phase of data preparation, the first results were brought out through
a data analysis.

4.2 Data Analysis

The data analysis starts from the pre-elaborated data set obtained from the previous
data preparation phase and in particular eight types of analysis were made:

1. Route Analysis;

2. Driving Mode Analysis;

3. Circuit Analysis;

4. Temporal Frequency Analysis;

5. Nearby Service Shop Analysis;

6. Fleet Analysis;

7. Weather Analysis;

8. Country Comparison Analysis.

4.2.1 Route Analysis

The first analysis focuses on understanding what kind of places and streets are
typically frequented by a Lamborghini customer. This analysis has been studied by
correlating the GPS position and the powertrain data of the vehicle, such as speed
and acceleration.

36
 Figure 4.3: Road and places most frequented

In fact, as can be seen in the picture 4.3, the percentages of the graphs refer to the
total time, therefore the correlations depend on the time in which a determinate
condition is observed, for example if 130 Km/h have been recorded for half an hour,
probably the car has been driven on the highway. So by correlating vehicle data,
GPS and duration time, on the Dashboard it is possible to see the most places and
streets frequented by Lamborghini customers.

4.2.2 Driving Mode Analysis

In this second analysis the result was the characterization and clustering of the
most used driving modes. In this specific case, since we are working with data on
Huracans, the following three clusters have been considered for simplicity: Strada,
Sport, Corsa. This would not be quite correct as for example the Huracan STO
model has different driving modes which are: STO, TROFEO and PIOGGIA. The
same would also apply if the analysis were extended to completely different models
such as the Urus and Aventador, which would have different driving modes to take
into account; but the choice of dividing the driving modes into Strada, Sport and
Corsa was made to simplify things because the number of STOs considered within
the data-set was negligible compared to the number of Spyder EVOs and Coupe
EVOs, but also because for the purposes of the PoC this was not strictly necessary
but the most important thing was to extract value from all these analyses.

37
 Figure 4.4: Driving mode usage by country

For example, in the graph 4.4, you can understand how the three driving modes are
used between different countries, it can be intuitively noted that the Corsa setting
is the least used in percentage compared to the other two.
Similar analysis was also made between the different Huracan models: Spyder EVO,
STO Coupe and Coupe EVO. What emerges in this case, 4.5, is that depending on
the model chosen, the use of the single driving mode varies a lot. For example, who
buys a Spyder EVO uses much more than the others the Sport mode.

Figure 4.5: Driving mode clusters per model

Finally, it was also studied on the entire fleet of vehicles analyzed what is the use
of these three drivign modes. As can be seen from the graph 4.6, the driving mode
most used by Lamborghini customers is the Sport mode and the least used is the
Corsa mode.

38
 Figure 4.6: Driving mode usage

Obviously, these analyses are a great cue to start thinking about why some driving
modes are more used than others, leaving ample space to new alternatives that the
company could introduce on the base of these studies.

4.2.3 Circuit Analysis

The third analysis was all focused on the study of driving habits when customers
go with their car on the track. In this case, it should be specified that only Italian
circuits have been considered; in fact, the GPS data of the circuits were not part of
the static data-set, but these have been added externally for the purpose of detecting
interesting use cases.
First of all, in this case it was analyzed which were the circuits most frequented by
customers and, as can be seen from the graph below 4.7, for each circuit we went to
see how many were the total visits and how many were the visits of each VIN, so
that we could have a more precise estimate of how many different people attended
that particular circuit. The study has shown that the two circuits most frequented
by Lamborghini customers are the Franciacorta and Misano circuit.

39
 Figure 4.7: Most visited circuits in Italy

After this, the monthly visits to the circuit of Italy were analyzed, and what was
discovered, as was easy to imagine, that during the winter months virtually no one
brings his Lamborghini on track, but prefers the summer/autumn period (figure
4.8).

Figure 4.8: Number of visits to circuits by month

Another very interesting estimate that has been made is how much time passes before
a customer brings their Lamborghini on a circuit (figure 4.9). What has emerged
is that on average a customer, after fifty days of purchase, makes his first visit to a
circuit, and brings the car back to the circuit after an average of thirty-three days.
This is a very interesting fact becaus, after having done the first experience on the
track, the time that passes with the second visit is much less than it took the first
time.

40
 Figure 4.9: How long does it take between a visit to a circuit and another?

Finally, we also analyzed the number of unique VIN to the circuits for each visit
number; and the thing that stands out most from this study is the fact that only one
VIN out of 21 totals came to bring his car seven times on track, so here the question
that comes naturally to ask is: why customers do not have all this desire to go on
track frequently? These are excellent points for reflection on which Lamborghini
will have to reflect in the future.

Figure 4.10: Number of unique VIN to the circuit for each visit number

Of course, it is only fair to point out that all the analyses in the current chapter are
based on data that were recorded from January to November 2021, i.e. just under
a year, so these are qualitative analyses; it is difficult to say that the driving habits
of a Lamborghini customer are always these, but it would be incorrect because we
have analysed far too small a time span to draw conclusions.

41
4.2.4 Temporal and Frequency Analysis
In this type of analysis we went to understand how cars are frequently used over time
during the year and here several ideas emerged. For example, as shown by the figure
4.11, the months when on average the cars are used more by their customers is in
the hottest months and the trend is that a Lamborghini Huracan is used more at the
weekend than during the week. So, already from this first analysis, the Lamborghini
in the eyes of its customers, it does not turn out to be a car to be used every day
to go to work but more a car to be used in the free moments.

Figure 4.11: Monthly and weekly usage of cars

In addition, it was also seen when typically a journey begins and what has emerged
is that those who have a Lamborghini Huracan tend to begin its journey mainly
in the late afternoon, from 4:00 PM onwards. This result verifies even more the
analyses made previously, that this type of Lamborghini model is more used in the
free moments of the week rather than like a daily car.

Figure 4.12: At what time a journey usually begins?

4.2.5 Nearby Service Shop Analysis

In this study it was analyzed how many times a customer Lamborghini is near the
service shop but we must make a clarification because otherwise this would lead to
wrong conclusions. When a customer is near a service shop it is not necessarily said
that he/she has had a problem with the car and therefore needs for assistance, but
simply because the whole analysis is GPS based and thus recognize the fact that
a VIN is in the vicinity of a service shop. Moreover, also in this case, as for the
circuit, only Italian service shops were considered.

42
First of all, we saw immediately the number of visits per service shop (figure 4.13),
both total and the number of visits of the unique VIN. What is easy to notice is
that the most frequented service shop is that of Milan, especially the 112 total visits
were generated by only 20 vehicles, which means that the same vehicle had to deal
with the service shop several times so here it would be interesting to understand
the reasons for all the other visits made, if they passed there by accident or had
problems with the car several times.

Figure 4.13: Total number of visits per service shops

In addition, we also calculated the number of visits to the service shop distributed
over time (figure 4.14); the good thing is that the result is very consistent with the
analysis made previously because before it was pointed out that the period in which
a Lamborghini Huracan is mostly used is summer/ autumn, and this also coincides
with the greater number of visits to the service shops, this shows consistency with
the previous one analyses.

43
 Figure 4.14: Number of visits to service shops by time

Another very interesting analysis that has been done is how much time passes be-
tween a visit and the following at the service shop (figure 4.15). In this case, the
most noticeable thing is that little time elapses between two visits and also here it
would be interesting to understand the reason, if the customers constantly encounter
problems with the car or they are visits untied from all this.

Figure 4.15: How long does it take between a visit to the service shop and another?

This analysis regarding the service shops has great potential if carried out in more
detailed way, such as correlating this GPS-based analysis with the recorded data of
the service shop, so that it can be seen very precisely whether the customer actually
went there because of problems with the car, or whether he/she just happened to
be there. Obviously understanding the real reason why a customer is near a service
shop would bring more value to this use case and this would mean increase the value
of the company, allowing to position itself in a position of advantage compare to its
competitors.

44
4.2.6 Fleet Analysis
This type of analysis is a more generic study than those made previously, includes
all the connected fleet, so what has been examined are more ”static” information
in the sense that are more correlated to the vehicles itself than to the driver. For
example, in the graph below, we calculated the kilometers travelled by the entire
fleet, the total hours that were in motion, the total countries in which they travelled
and how many vehicles in the fleet were connected during the period considered
from January to November 2021.

Figure 4.16: The general values attributed to the fleet

In addition, we saw how in the time-frame considered (figure 4.17), i.e. January to
November, what was the distribution of connected vehicles over time. This may be
a rather trivial indication, but it could come in handy if we come across connectivity
problems; for example, if a certain vehicle is connected, it must be present in the
dashboard; otherwise, this is a problem that needs to be resolved as soon as possible.
As seen from the graph, selling connected vehicles is now a MUST, and taking care
and especially monitoring the connectivity of the customers in the right way shows
how Lamborghini is becoming more and more connected and how much it is investing
in this direction.

Figure 4.17: Number of active connected cars per moth

Finally, a study was also made on the tons of CO2 produced by the fleet of Lam-
borghini connected (figure 4.18). In this specific case the estimate of CO2 produced
was calculated using some external parameters, such as the consumption of each Hu-
racan model, so by making a correlation between kilometers traveled, type of model
considered and consumption for each model, we are able to visualize the graph of

45
CO2 produced. This graph will have to be made more accurate in the future, i.e.
the estimate of the CO2 produced will have to be based directly on reading the tank
level indicator in order to have a direct correlation on how much CO2 the machine
is emitting. This graph will have to be made more accurate in the future, i.e. the
estimate of the CO2 produced will have to be based directly on reading the tank
level indicator in order to have a direct correlation on how much CO2 the machine
is emitting. This will be a cue for the next steps to be taken in order to make the
data we read on the dashboard more and more valuable.

Figure 4.18: Tons of CO2 produced per month

4.2.7 Weather Analysis

This analysis has been done in such a way as to get out of the interesting use cases, in
order to consider also the driving habits on the base of the meteorological conditions.
As mentioned above, for simplicity, only three cities have been considered, since most
of the connected cars taken into account are located in these three areas: Milan,
Miami and Los Angeles. In fact, as you can see in the figure 4.19, it has been related
to the weather conditions that there were in those three cities during the year with
the use of the car, so as to verify whether the car was used more or less under
certain conditions. What was highlighted is that, for example, in Miami last year it
rained heavily about 20% of the time but the car, in those conditions of heavy rain,
was used 23%, which means that, despite little rain in Miami, the vehicle is used
regardless of light or heavy rain conditions.

46
 Figure 4.19: Weather conditions of use of the vehicle by city

Another interesting analysis was to relate driving modes to weather conditions. This
pointed out that as the intensity of the rain increases, the use of the more aggressive
trim, Corsa, decreases; conversely, the more intermediate mode, the Sport mode,
increases with the intensity of the rain.

Figure 4.20: Driving mode distribution by weather conditions

The thing that emerges is that, even in heavy rain conditions, the preferred mode
is that of Sport and it would be interesting to go and understand why.
Knowing and investigating why customers make certain choices could bring a consid-
erable advantage to the company, allowing it to realise and implement increasingly
customer-driven strategies so that the driving experience can be more and more
unique.

47
4.2.8 Country Comparison Analysis
Finally, the last analysis made is to put in comparison the driving habits between
different countries. The comparison in this case was made between Italy and Amer-
ica but other countries could also be compared. Again, since most of the data were
for Italian and American vehicles, for convenience the comparison was made directly
between Italy and America, but later when the data-set gets bigger and bigger will
be possible to compare any country we want.
The first thing that has been highlighted from the graph figure 4.21 is that the car is
used more by Italian customers during the year than the Americans, and at the same
time, the use of a Lamborghini Huracan also in America, as in Italy, is prevalent
over the weekend, this also emphasizes the previous analysis already made, when it
was said that a Huracan is used expressly only in leisure time.

Figure 4.21: Vehicle usage divided by country

Another thing that turned out to be slightly different is the use of driving modes. In
particular, in America there is a propensity to use the vehicle in a somewhat more
aggressive mode than in Italy, in fact the Strada setting is little used by Americans.

Figure 4.22: Driving mode usage by country

Finally, another interesting analysis is about the time when a journey begins. Here
the difference is clear, in America a Huracan is typically used to go out in the evening
until late at night, in fact as is seen in the figure 4.23, an American journey ends

48
around 4:00 AM; in Italy instead the trend is the opposite, it is preferred to use the
car at more convenient times, without making a worldly life.

Figure 4.23: At what time a journey usually begins?

All the analyses just seen allow the first insights on how Lamborghini customers
like to use their Huracan and this brings great value to the company itself which
can already understand, based on these registered driving habits, on which driving
services it must more push in the next strategies.

4.3 Privacy and data protection in connected ve-

hicles
Not only drivers and passengers, but also vehicles are becoming more and more
connected, which is why, also in the automotive sector, there is more and more talk
about privacy and data protection on board.
Many models launched on the market in recent years from all the most important
OEMs, integrate sensors and connected multimedia equipment that can collect and
record various parameters such as engine performance, driving habits, places vis-
ited and potentially even biometric data for authentication or identification and/or
safety (i.e. the physical safety of people) purposes. This data processing takes place
in a complex ecosystem, which is not limited to the traditional players in the au-
tomotive industry, but is also influenced by the emergence of new players from the
digital economy.
These new service providers may offer infotainment services such as online music,
road conditions and traffic information or provide driver assistance systems and ser-
vices, vehicle condition updates, usage-based insurance or dynamic mapping.
Furthermore, as vehicles are connected via communication networks, road infras-
tructure managers and telecommunication operators involved in this process also
play an important role with respect to the processing of personal data of drivers
and passengers (think of future 5G networks that will increase connectivity from
the vehicle to the outside, i.e. so-called V2X). Increasingly connected vehicles gen-
erate increasing amounts of data, most of which can be considered personal data as
it relates to drivers or passengers. Even if the data collected by a connected car is
not directly related to a name, but to technical aspects and characteristics of the
vehicle, it will always affect the driver or passengers of the car. For example, driving
style data such as:
• distance travelled;

49
 • wear and tear on vehicle parts;

• cameras that study the driver’s behaviour, as well as information on other

people who may be inside or outside the vehicle.

4.3.1 Privacy and data protection in connected

vehicles: the principles
In the field of cyber security, the three main objectives are well known and are often
referred to according to the CIA triad (Confidentiality, Integrity, Availability), which
we take up here as a definition:

• confidentiality: information is not made available or disclosed to unautho-

rised individuals, entities or processes;

• integrity: data cannot be altered or deleted in an unauthorised or undetected

manner;

• availability: the computer systems used to store and process information

must function properly and be available as resources at all times.

Similarly, the following three objectives can be defined at the level of privacy:

• predictability: The ability to know what data from the various devices
around us is being collected and how the information will be processed and
handled thereafter;

• manageability: Correlation between data can lead to sensitive information,

we need to know the granularity of how our data is processed and how to
change or delete it;

• non-associability: i.e. treating data in such a way that it cannot be associ-

ated with individuals, avoiding tracking and profiling.

This ecosystem is not very different in the end from the IoT (Internet of Things)
world where the challenges in terms of privacy appear very similar. Today, in fact,
a connected vehicle is already a network element capable of ’talking’ with other
vehicles, with roadside systems, with servers in the cloud and, of course, with the
passengers themselves, who increasingly interact with on-board functionalities via
Bluetooth or Wi-Fi or USB/SD ports (such as memory or connectivity devices).

4.3.2 Privacy and data protection in connected

vehicles: Regulations
When it comes to privacy and data protection in connected vehicles, it is also pos-
sible to comply with a number of regulations and best practices. On 28 January
2020, the EDPB (European Data Protection Board) issued the European guidelines:
’Guidelines 1/2020 on processing personal data in the context of connected vehicles
and mobility-related applications’.

50
This document focuses in particular on the processing of personal data in connection
with the non-professional use of connected vehicles: e.g. drivers, passengers, vehicle
owners, hirers and so on. More specifically, it deals with personal data:
• processed inside the vehicle

• exchanged between the vehicle and personal devices connected to it (e.g. the
user’s smartphone);

• collected inside the vehicle and exported to external entities (e.g. vehicle
manufacturers, infrastructure managers, insurance companies).
The EDPB document summarises the guiding principles for the protection of per-
sonal data:
• importance and minimisation of data;

• data protection by design and by default (privacy by design and by default);

• processing of personal data;

• anonymisation and pseudonymisation;

• data protection impact assessment (DPIA);

• data subject rights;

• security and confidentiality;

• transfer of personal data to third parties;

• transfer of personal data outside the EU/EEA;

• use of in-vehicle Wi-Fi technologies.

The main conclusion to be reached through reading the document is regarding ve-
hicle drivers and passengers who may not always be adequately informed about the
data processing that takes place in or through a connected vehicle.

4.3.3 How Privacy was addressed for PoC pur-

poses
In all of the PoCs seen in the previous chapters, being PoCs for internal and not
external purposes, Lamborghini decided not to immediately address in a structured
way the privacy issue, but since the sole purpose of the work was to make the
company understand how much potential there was behind the data of the connected
Huracans, we only focused on how to get the most value out of the data-set.
The only thing that has been done is to consider all available data as anonymised
aggregated data, so that there was always maximum confidentiality with regard to
each VIN (Vehicle Identification Number) that was used for the analysis in question.
Since this PoC was only disseminated internally within the company for the purpose
of explaining the PoC, a proper PIA (Privacy Impact Assessment) was not necessary,

51
which, however, will have to be done if the company decides to bring the Big Data
project into production; in this case, the data collection must be done accordingly
with Privacy and Security regulations.

52
Chapter 5

Lamborghini KPI

The acronym KPI stands for Key Performance Indicators and these are a set of
quantifiable measures that a company uses to evaluate its performance over time.
KPI are key to helping you make informed and informed business decisions based
on the performance of the activities being analyzed.
In the previous chapter, many different types of analysis were done and what we
need to focus on now is if Lamborghini can use some of these analyses as KPI that
can help the company make strategic decisions. In my opinion, there are several very
interesting metrics seen in the previous chapter, that can be taken in the future as
KPI on which to start and base interesting business discussions and now we describe
the main ones.

5.1 Driving Mode KPI

When we previously covered all the analysis about driving modes, very interesting
results emerged. For example, based on the type of Huracan model bought, the
use of driving modes was different. In fact, as can be seen from the driving mode
clustering graph (figure 5.1), the CORSA mode is used very little compared to the
other two available modes. This could be a first starting point from which to start
thinking about whether to continue spending energy and money on the production
of this driving mode or not. For example, for the Huracan STO model, which was
created precisely for track use, it would make sense to maintain the use of the Corsa
driving mode, which in reality, as already mentioned in previous chapters, within
the STO the driving modes do not take these names but are: STO, TROFEO and
PIOGGIA; but in any case, it makes sense for this one to maintain a driving mode
based on the Corsa concept. On the other hand, for the Huracan Spyder EVO and
Coupe EVO models, which are perhaps less suitable cars for the track, one could
probably think here of replacing the Corsa driving mode with another one which
might be appreciated by customers who own a Huracan EVO.

53
 Figure 5.1: Driving mode clusters could be used as KPI

Also, going to clustering the vehicles under consideration, the easy thing to notice
is that the Corsa driving mode is used very little, only a 5%. This data can also
be taken as a strong enough indicator for the company to reason about why the
customer uses that driving mode so little. Maybe customers feel unsafe driving and
prefer a less aggressive setup? Well, in this case, Lamborghini could promote to
these customers, through the use of the UNICA App, some ad-hoc driving lessons
with professional drivers so that they can feel more confident to use the Corsa setup
more frequently and bring their Lamborghini to the limit.
Instead, if the usage of Corsa setting always remains so low, Lamborghini could also
consider replacing it with another driving mode that could be used more by the
customer, or remove it completely without replacing it, thus saving a lot of money
in the production phase.

Figure 5.2: Driving mode variation clusters could be used as KPI

Another indicator worth thinking about is the variation in driving modes over time.
Indeed, the graph shows us (figure 5.2), on average over time, how much a driver
changes driving mode while driving the car, and this yielded interesting results.
In fact, irrespective of the Huracan models, most drivers prefer to drive using two
driving modes, while those who typically use all three or only one of the three
available are a very small percentage.
Obviously, we must always remember that these results emerged from the study of
a data-set that is based on a very small time interval of observation so we can not

54
directly draw conclusions, but what came out is that most of the customers who
own a Huracan prefer to drive using two driving modes. From the outcome that
emerged, so that we can go to stimulate the customer to use more frequently all
three driving modes present, always through the use of the UNICA App, it could
be possible to create rewards based on how many kilometers the customer drive
with a certain driving mode. For example, once he/she have covered a certain
mileage with the STRADA driving mode, a notification will arrive to the customer
telling him to go to his dealer to pick up a customized gadget such as a bracelet
or keychain; instead, if the customer had to do a certain mileage with the SPORT
driving mode,there will be the possibility to unlock a certain number of months free
within which the customer can take advantage of some On-Demand services that
typically Lamborghini makes available for a fee. Instead, when the customer will
exceed a certain mileage with the CORSA mode, then here a reward could be free
access to a certain circuit that is located near a location where the customer prefers
to go.
Obviously this ”reward game” will have to renew as the customer begins to unlock
rewards; then once the customer has reached the first reward with the use of a
driving mode, will be unlocked the achievement of the second reward with the use
of the same driving mode. This will bring great benefits to both the customer and
the company itself, because in this way the customer will be encouraged to drive
more frequently their Huracan and this will bring a great benefit also to the company
because it would mean storing more data to analyze and nowadays to own more data
means to have the chance to make more money.

Figure 5.3: Driving mode variation by country could be used as KPI

The last thing we need to consider as an indicator for driving fashions are using
these from country to country. In fact, as can be seen in the graph 5.3, in every
country the use of driving modes changes a lot, for example passing from Italy

55
to Austria, despite being neighboring countries, the use of driving mode STRADA
changes a lot; obviously the physiognomy and the regulations of a Country influence
a lot also on the driving style. For this reason, Lamborghini will have to carry out
an analysis of driving modes and customize them from country to country. For
example, if in Netherlands it is almost always used the STRADA driving mode, it
would be useless in this case to promote an event on the track, but maybe it would
be better to promote some Peace-Of-Mind service so that the driving experience is
as relaxing as possible; while in Austria, where the landscape allows the car to be
driven a little more to the limit of its performance, in fact the CORSA driving mode
is widely used, then here it would make more sense to promote a driving lesson with
a professional driver.
In general, the indicators seen so far, manifest the need to create OTA (Over-
The-Air) Ego driving modes, i.e. driving modes for the user’s personal use and
downloadable from the UNICA App, so that, regardless of the car model and the
customer’s driving ability, the user can at any time download and use the driving
mode that suits him or her best, so that it is as customizable as possible.
So taking this data as indicators, it would lead to the emergence of new marketing
strategies.

56
5.2 Circuit KPI
Also with regard to the analysis made on the most frequented Italian circuits, sev-
eral very interesting factors had emerged to be taken into account. In particular,
something that had been noticed that very few people just buy a Huracan tend to
take it straight to the track maybe they go there a couple of times and then they stop
altogether. Of course, to draw that conclusion as true, we would have to consider
more data than we have analyzed for this POC and a much larger time frame, but
for the moment we try to make consideration on this result that has emerged out
of this first analysis. Viewing and monitoring this type of data could be crucial to
understand on average if there is really great interest from the customer to attend
the track more assiduously and, based on the result, this could be a great indicator
for make Lamborghini understand how to move or how to improve certain strategic
choices.

Figure 5.4: Number of unique VIN to the circuit used as KPI

This can lead to asking the company why customers tend to take their cars to the
track less and less over time; maybe they don’t feel comfortable taking the car to
the extreme on a circuit. So these values could be kept in consideration, leading
the company to make different decisions, such as introducing on-demand services
more related to the circuit than to the road, or services aimed at rewards and one
of these could be a free day at a circuit or a daily lesson with a professional driver;
another option could be to make the vehicle a game console on which the customer
can simulate, through a projection on the windshield, a real experience on the track,
this will allow each customer to train before having a real experience on the track,
but at the same time it will allow to create a certain link between the car and the
driver because the simulation will take place inside the cockpit then using the pedals
and the steering wheel of the car itself. Another way to encourage more and more
customers to take to the track with their car is to implement a semi-autonomous
guide system inside the vehicle that will take the control of the whole vehicle only
when the driver makes mistakes while driving, such as cornering or braking before
the turn, this will generate greater safety to the driver, so it will also give the

57
possibility to those who are not a professional driver to be able to bring the vehicle
to the limits.
In general, before coming to any conclusions, the analysis has to be done in more
detail because typically spring and summer are the periods when cars are most often
taken to the track, so if a customer buys a car in winter it is not that he does not
want to go to the track, but simply because he waits for the warmer periods to go.
So the whole analysis must also be done taking into account the period of year in
order to draw more precise conclusions. Therefore, there are many ways to increase
the presence of Lamborghini customers on the track.

5.3 Nearby Service Shop KPI

Another very interesting analysis was that about the proximity of the service shops
and it was noticed that typically, on average, every 20/25 days a Lamborghini car
was in the vicinity of an italian service shop, the reason is not known, if the customer
is there by accident or to actually request a maintenance of the car for problems
related to the vehicle. But the interesting thing that could be studied is the possible
relationship that could be among the most used driving modes with respect to all
those cars that are registered near a service shop.
Now in this case the cars that are most often found near a service shop are the same
ones that have the most use in the SPORT and CORSA driving modes (figure 5.5);
this makes a lot of sense because the more the customer uses the track mode, more
the car will need maintenance, as the brakes, tyres and suspension of the car are put
under more strain. So it could be taken as an indicator the relationship that there
is between how much the Corsa/Sport mode was actually used by that cars that
have been seen near a service shop, because maybe a possible scenario for the future
could be that those who use these two driving modes more frequently, through the
predictive maintenance, a direct relationship can be established between customer
and service shop, so that the customer can be notified a reminder for the next
maintenance, or perhaps be alerted as soon as something strange is detected by the
service shop, such as strange engine-related values, or when parts are available for
replacement, as for example replacing the suspensions with other more performing
in order to reduce vibration.

58
Figure 5.5: Relation between driving mode most used by the vehicles that have been
seen near a service shop as KPI

Predictive maintenance allows optimisation of all maintenance operations with sig-

nificant benefits in terms of cost reduction and increased productivity. A predictive
maintenance system, which can be customised according to requirements, can au-
tomate most maintenance management processes, with considerable savings in time
and resources. Intervening only if and when needed leads in fact to a reduction
in machine downtimes, labour costs as well as machine replacement costs, with a
consequent increase in production continuity and revenue, but at the same time also
to greater customer loyalty due to excellent customer care.
Obviously Lamborghini is also focusing and working on the predicitive maintenance
side in order to take advantage of this additional service as soon as possible.

5.4 Weather KPI

Another interesting analysis was that relating to the use of vehicles on the basis of
weather conditions. In that case, the study had only been conducted for three cities
but, despite this, was always found something interesting. For example, it was seen
that in Milan, despite having rained heavily 20% of the time during the year, the
use of the vehicle in these conditions was used 23% of the time. This means that,
despite the weather conditions were not optimal, the Lamborghini customers like to
go out with their Huracan in these circumstances.

59
Figure 5.6: Relation between the usage of the vehicle and the weather condition as KPI

Therefore, taking this as an indicator, the company could start thinking about
developing safety services in heavy rain, such as a semi-assisted driving in case the
driver loses control on the wet road, so that even those customers who may not yet
feel very confident with the machine can feel even more comfortable and secure.
Furthermore, since it has just been noted that the Huracan is also used a lot in the
rain, the RAIN drive mode could also be brought into the Coupe EVO and Spyder
EVO models, since the STO already has it inside, in order to make the driving
experience even more complete and fascinating.
Again, as in the case of the driving modes seen above, it would be interesting as a
solution to have OTA driving modes, so it would be possible via UNICA App to
download and use driving modes that the car itself does not possess at the production
stage, so that the user in whatever circumstances he finds himself, can use the driving
mode most comfortable to him.

60
5.5 Country Comparison KPI
Finally, the last analysis made was about the country comparison, how the vehicles
are used depending on the country. This study is in general a source of great value
because it allows to understand what are the main driving habits depending on the
country of origin, thus allowing to develop ad-hoc products and services based on
their habits.
For example, we analyzed in the previous chapter the time at which, usually, begins
a journey in Italy and America. The result was quite contrasting, as it found that
in America the car is used purely at night, even until 4 in the morning; while in
Italy the use is different, you prefer to go around at more convenient times.

Figure 5.7: Country comparison among at what time a journey usually begins as KPI

Just going to study the greater differences of use that there are between a country
and the other, it would allow the company to focus on different aspects on the basis
of the market considered; in fact as for America maybe it could be more reasonable
thinking more on services for nightlife than Italy in which the addition of these
services could be avoided.
At the same time, once we understand the most widespread habits within a country,
we may also use this information for digital marketing purposes, as for example the
organisation of events can also be influenced by these analyses, e.g. in America it
would make more sense to organise a Lamborghini event in the night, whereas in
Italy in the afternoon.

5.6 Electrification KPI

Finally, through the merger of different sources, it would be possible to obtain some
very useful results for the electrification that as never is a very topical theme, in
fact as already said by the European Commission from 2035 will have to stop the
sale of ICE vehicles; for that reason, it would be useful to have indicators that can
help in view of the change that is coming.
In fact, as seen in the previous chapter, we have several information that could be
useful to us, as for example we had seen how long a journey lasted on average of
a customer who owns a Huracan, calculated from the moment the vehicle is set in
motion until the moment of switching off (figure 5.8).

61
 Figure 5.8: Average duration of a journey used as KPI

In addition, it would also be possible to read the signals of the fuel level sensor and
relate this to the distance travelled by the individual vehicle, in order to arrive at
an accurate estimate of consumption.
All this information, if related to each other, could prove to be excellent indicators
and information that would be very useful in the design or choice of electric batteries
that will have to be introduced in the Lamborghini of the future. This will open the
door to new strategic decisions for the company that could also bring great economic
savings, because for example if the distance estimated for each journey is around 25
Km, it would be useless to think to put inside the Lamborghini a battery with an
autonomy of 1000 Km, but maybe it would make more sense to put smaller batteries
that would bring great advantages both in the management of space and also from
an economic point of view, but, at the same time, this would not affect the customer
experience because in any case the customer will be able to drive without any kind
of worry.

62
Chapter 6

Next Steps

In light of the above, the entire thesis work that has concerned me in first person
has been the PoC discussed in detail in chapter four. The entire analysis has high-
lighted possible value KPIs that could lead Lamborghini to take new strategies in
the future. All the analyses and considerations have shown great potential, leaving
all the members of the strategy departments very satisfied. For this reason, there
was a planning on what could be the next steps to be done, in order to add even
more value to the work that has already been done. The result was that there would
be many different things to be able to expand and add, but each of these has also
highlighted different temporal needs, for this reason, the next future steps can be
divided into three macro-areas:

1. Short Term

2. Middle Term

3. Long Term

6.1 Next Steps: Short Term

In this section will be shown all those next steps that could be implemented in the
short time, which have a very low commitment and that could be brought to a short
term. The short term next steps are as follows:

• For the PoC we have always considered a static database, now the next step
will be create a real time dashboard directly linked to SDP database;

• Extend the circuit analysis to other countries in order to make a more accurate
study and see if also in other parts of the world the relationship that the
customer has with the track is the same or different from the one seen in small
part by us;

• Specify the number of vehicles considered on total and on each country, because
for the moment it is not specified on the total of the vehicles taken into account
how many belong to a particular country and it would be useful to have this
information to go more and more customizing the driving habits country by
country;

63
 • Further comprehension of the car in racetracks; for the moment we have too
little data and a time frame of observation too small to deduce conclusion
about the real behavior of the user on the track. Until today we have seen
that typically in Italy the presence on the track is not too crowded, but in fact
we should collect more data to deepen the study of the circuit;

• Improvement of the platform architecture considering the continuous data

growth; so the future step will be to consider a Cloud architecture that will be
more robust than the one used in the PoC, at the same time it will also have
to be more secure because it will have to handle a huge amount of data whose
privacy and security will be very important, and finally, it will have to be fast
enough in the distribution and computation of data;

• The analysis of the driving modes will have to be done in more detail, i.e.
for each Huracan model considered, it will be important to use the driving
modes that really belong to it, in fact for the STO we have already said
that the driving modes are: STO, TROFEO and PIOGGIA. Consequently,
the clusters can no longer be STRADA-SPORT-CORSA, but the three new
clusters referring to the corresponding driving modes must also be considered
for STO.

6.2 Next Steps: Middle Term

In this section we will see what may be other next steps that will not be immediate
but it will take some time before seeing them, for this reason they are considered in
the middle term and are as follows:

• Get access to customer database in order to link the information and have a
complete data-set client-car;

• Investigate possible application for the next gen SSC (Super Sport Car);

• It would be interesting also to see how many vehicles frequently pass through
those urban areas called Zero Emission Zones, in which it is possible to drive
only in electric mode, and analyze how many kilometres are travelled in full
electric mode. In that way we could understand the mileage allowed to esti-
mate what kind of batteries need to be put on board so that the customer can
travel electrically inside these zones without the risk of having little autonomy;

• Privacy On/Off in order to see in real time the privacy mode, to have every-
thing automated and easy to trace;

• U.S.A study based on zone/region, in order to differentiate and cluster better

the driving habits as the zone changes;

• In the PoC, the calculation of CO2 emissions was based on the kilometres
travelled multiplied by the number of vehicles considered and in turn multiplied
by the value of CO2 emitted for each Huracan model (calculated as g/Km).
The next step would be to go directly to read the signal, which travels on

64
 CAN, based on tank emptying, so as to have the most accurate estimate of
CO2 emitted;

• Monitoring of the Remote APP services;

• For the moment, the weather analysis has only been implemented for three
cities: Milan-Miami-Los Angeles; perhaps in the future, as next step, the
product marketing will need to expand the same analysis for other cities that
are relevant for Lamborghini commercial strategies.

6.3 Next Steps: Long Term

In this section, instead, we will go to consider all the next steps more complex that
will have to wait a long time before seeing them realized. The next steps are as
follows:
• Consider in the future also the test vehicles behavior that have been always
excluded in the previous PoC;

• Try to read the passenger sensor signal in order to understand if the driver is
alone in the car or together with another person. This might be interesting to
understand whether Huracans are typically used only by the individual driver
or whether they are used together with a passenger;

• Link between Service Shop and LIASS (Lamborghini Integrated After Sales
System) in order to have a better knowledge if the driver really brought his car
to the service shop for a maintenance or if instead he/she only passed there
by accident;

• Development of the use case for predictive maintenance, in particular analyze

the vehicle data, such as the brake pads, in order to identify the point of fault
and inform the user through the App UNICA before the breakdown occurs;

• Integration of the analysis processes on Urus fleets in order to expand the Big
Data study also on the other Lamborghini product lines;

• DTC (Diagnostic Trouble Codes) acquisition in order to improve the reliability

of each components for better diagnostic analysis;

• Increase the frequency data acquisition, today each data item is acquired every
ten seconds, in the future it would be good to improve this value. The reason
is that because if we acquire each data every 10 seconds the analysis we could
do would be poor from the point of accuracy because a Lamborghini goes from
0-100 km/h in about 3 seconds, if we acquire every 10 seconds we will see very
little correlated data between one and the other. So for values relating to the
powertrain or vehicle dynamics it would be better to have very high sampling
rates, even under a second; while for readings of other things like the door or
the passenger signal it would be useless to acquire them every second because
they are not things that vary every second, so in this case it would make sense
to keep the sampling rates lower.

65
Finally, these are all the next steps that have been set to improve the previously
treated PoC. These new ideas will be used both to adjust the analysis already made
but also to improve and expand the entire study, so that all the analysis can be as
accurate as possible to bring us to the KPIs of great value.

66
Chapter 7

Conclusion

Automotive connectivity is changing faster than ever, significantly increasing the

potential for data monetization for players across the ecosystem. Data suppliers,
such as OEMs and vehicle fleets, are well positioned to benefit, as are insurance
players, companies in the automotive aftermarket, cities, infrastructure providers,
and other data customers. Importantly, all stakeholders must act fast. Given the
industry’s current underperformance on data monetization, new players with inno-
vative approaches could rapidly gain an advantage over slower-moving incumbents.
Those that fail to act now will miss the opportunity to differentiate themselves in
one of the industry’s key customer-facing spaces. While OEMs, suppliers, and other
players along the value chain increasingly realize this imperative, they have not yet
consistently created new offers and services that customers find compelling. They
often fall short because customer expectations keep increasing and technology ad-
vances are occurring rapidly. If they continue to underperform, their brand appeal
and profit pools could suffer, decreasing their market share. Conversely, those that
harness the opportunity before them may unlock new profit pools for the industry
and enable new, profitable growth.
Lamborghini has been one of the first OEMs, in the field of super sport-cars and
luxury, to move towards this new trend, all the projects mentioned in previous chap-
ters and the same thesis work are great witnesses.
In summary, in this thesis work the main objective was to analyze the data of con-
nected cars, going to understand if these could really achieve important and valuable
results. Lamborghini is now a year that is increasingly trying to meet the demands
of the market and especially the trend regarding Big Data. Precisely, because of the
new trend in the data of connected cars, Lamborghini has carried out four different
initiatives in the last year with the development of four PoCs. The goal, in each
case, was to be able to get the most value from the data analyzed in order to obtain
some KPIs that can direct future business models and decisions. In particular, my
contribution has not been linked to all four of the PoCs discussed but only to one
of these, on which I had different decision roles, as:

• What types of use cases to focus on;

• How to try to correlate data to get interesting use cases;

• Dashboard layout and graphical view;

67
 • Notify some possible problems related to the analysis made and solve them
with possible next steps;

• Continuous contacts with the technicians working on the dashboard to guide

them in the choices made by the Strategy-Connectivity team.

The results obtained for the purposes of the PoC were very convincing, making open
the eyes to Lamborghini that great value is behind data of the connected cars and
how much it is worth investing in this new trend of Big Data because it is really full
of opportunities and in the future there will be more and more, given the continuous
increase of connected cars.
In conclusion, the PoC had as only objective to make to understand the potentialities
that have these data and the great benefits that can be drawn from it; considering
the obtained results and the opinion of the team, the objective has been reached
with great success. In fact, nowadays at Lamborghini is increasingly clear and aware
that to differentiate from its competitors and remain competitive in the market, this
is the right direction on which to continue to focus and on which we will have to
work more and more soon.

68
Bibliography

Deloitte (2021), Software is transforming the automotive world

Deloitte (March 2022), 2022 Global Automotive Consumer Study

Deloitte June 2020, Future of Sales and Aftersales, impact of current industry trends
on OEM revenues and profits until 2035

Frost Sullivan (2021), Top 5 Connected Car growth opportunities

Frost Sullivan (April 2021), What’s driving the 2021 Global Automotive Industry

GSMA, Connected Cars: Business Model Innovation

McKinsey&Company (Compendium 2019/2020), The future of mobility is at our

doorstep,

McKinsey&Company (January 2016), Automotive revolution - perspective towards

2030

McKinsey&Company (January 2019), RACE 2050 - A vision for the European au-
tomotive industry

McKinsey&Company (May 2022), Can the automotive industry scale fast enough?

McKinsey&Company (October 2020), Reimagining the auto industry’s future: It’s

now or never

PWC (2021), Five trends transforming the Automotive Industry

REUTERS (March 2022), Trends Transforming the Automotive Industry in Europe

Roland Berger (January 2020), Computer on wheels: Disruption in automotive elec-

tronics and semiconductors

SBD (2019), Data Monetization and strategies for Connected Car

SBD (2022), Connected Car Data

69