Big Data Analytics
Data Visualization
Motivation
Visual Perception
Eye Vs. Camera
Gestalt Principles
Context, Preattention, Magnitude Estimation
Visual Attributes and Visual Mapping
Evaluating Visualization
Effectiveness, Expressiveness, Integrity, Consistency
Chart Junk
Imdad ullah Khan
Imdad ullah Khan (LUMS) Data Visualization 1 / 101
Data Visualization
Data Visualization is the graphical representation of information and data.
Visual elements like charts, graphs, and maps, provide an accessible way
to see trends, outliers, and patterns in data
Enables the quick interpretation of data
Helps communicate information clearly and efficiently
Essential for data-driven decision making
The primary goal of data visualization is not only to present data but to
provide insights that are not immediately obvious through raw data.
Efficiency: Rapidly digest large amounts of data
Pattern Recognition: Identify patterns, relationships, and outliers
Storytelling: Translate findings into a narrative to influence
decision-making
Imdad ullah Khan (LUMS) Data Visualization 2 / 101
Data Visualization
What information consumes is rather obvious: it consumes the attention
of its recipients. Hence a wealth of information creates a poverty of
attention, and a need to allocate that attention efficiently among the
overabundance of information sources that might consume it.
Herb Simon, Scientific American, 1995
Data volume and velocity is much higher than our ability to draw
knowledge from it
Visualization helps draw knowledge form data (beyond statistical
inference)
Visualization reveal information that statistics may not
Visualization of scientific data magnifies the capabilities of science to
understand the universe
Imdad ullah Khan (LUMS) Data Visualization 3 / 101
Data Visualization: Why?
Imdad ullah Khan (LUMS) Data Visualization 4 / 101
Data Visualization: Why?
Bathymetric map
Nautical Chart
Reveal Hidden dangers
Helps marine navigation
Imdad ullah Khan (LUMS) Data Visualization 5 / 101
Data Visualization: Why?
Popular belief in 1850: cholera spreads via airborne transmission
Dr. John Snow plotted each death on a London map
Noticed clusters around a certain contaminated well
Steve Marschner @Cornell
Imdad ullah Khan (LUMS) Data Visualization 6 / 101
Data Visualization: Why?
Popular belief in 1850: cholera spreads via airborne transmission
Dr. John Snow plotted each death on a London map
Noticed clusters around a certain contaminated well
images source: https://www.theguardian.com/news/datablog/2013/mar/15/john-snow-cholera-map
Imdad ullah Khan (LUMS) Data Visualization 7 / 101
Data Visualization: Why?
States mean income and fraction of college degree holders
source: Bradley Hemminger, Uni. of North Carolina
Which state has the largest and the smallest —?
Which states are outliers if any?
How is income related to college degree?
Imdad ullah Khan (LUMS) Data Visualization 8 / 101
Data Visualization: Why?
Can easily tell what is largest/smallest in every dimension
Imdad ullah Khan (LUMS) Data Visualization 9 / 101
Data Visualization: Why?
Visualization helps identify relationship easily as compared to raw data
Imdad ullah Khan (LUMS) Data Visualization 10 / 101
Data Visualization: Why?
Outliers stand out and get identified easily
Imdad ullah Khan (LUMS) Data Visualization 11 / 101
Data Visualization: Why?
Anscombe’s Quartet: Four datasets with identical statistics
x 4 5 6 7 8 9 10 11 12 13 14
y 4.26 5.68 7.24 4.82 6.95 8.81 8.04 8.33 10.84 7.58 9.96
x 4 5 6 7 8 9 10 11 12 13 14
y 3.1 4.74 6.13 7.26 8.14 8.77 9.14 9.26 9.13 8.74 8.1
x 10 8 13 9 11 14 6 4 12 7 5
y 5.39 5.73 6.08 6.42 6.77 7.11 7.46 7.81 8.15 12.74 8.84
x 8 8 8 8 8 8 8 8 8 8 19
y 6.58 5.76 7.71 8.84 8.47 7.04 5.25 5.56 7.91 6.89 12.5
µx = 9 σx = 3.316 µy = 7.500 σy = 2.031
▷ Edward Tufte, The visual display of quantitative information
Imdad ullah Khan (LUMS) Data Visualization 12 / 101
Data Visualization: Why?
Anscombe’s Quartet: 4 datasets with identical regression line
Imdad ullah Khan (LUMS) Data Visualization 13 / 101
Data Visualization: Why?
The eye and the visual cortex of the brain form a massively parallel processor
that provides the highest-bandwidth channel into human cognitive centers.
Colin Ware, Information Visualization, 2004
Visual system is the highest bandwidth channel to the brain
70% of body’s sense receptors reside in our eyes
Metaphors to describe understanding often refer to vision (“I see,”
“insight,” “illumination”) ▷ Thinking with our Eyes
Need an efficient way to understand Big Data
Imdad ullah Khan (LUMS) Data Visualization 14 / 101
Data Visualization: Why?
Makes the vast amounts of data more comprehensible
Reveals invisible parts in data that we don’t have access to otherwise
Analyze things that are otherwise difficult
Allows for quick decisions based on real-time data visualizations
Capture events
See things at a level that is not available at our own perception
Magnifies our ability to understand things better
Help us tell a story
Visualizations transcend language barriers and are universally
understandable
Imdad ullah Khan (LUMS) Data Visualization 15 / 101
Visual Perception
Imdad ullah Khan (LUMS) Data Visualization 16 / 101
Visual Perception
Understanding how we perceive visual information is crucial for designing
effective data visualizations
Knowing how the brain would read visualization enhances design
▷ Know your audience
Understanding mechanisms of the visual processing system and using that
knowledge can result in improved displays
Having an idea of human perception and psychology helps in optimal
visual mapping and developing meaningful visualization
Imdad ullah Khan (LUMS) Data Visualization 17 / 101
The Visual Processing System
The human visual system is a complex mechanism evolved to process
information efficiently and effectively.
Eye as a Sensor: Captures light and transmits signals to the brain
Visual Cortex: Processes visual information to interpret shapes,
colors, and patterns
Cognitive Processing: Uses stored knowledge and context to make
sense of visual data
Imdad ullah Khan (LUMS) Data Visualization 18 / 101
Visual Perception: Eye vs Camera
Camera:
Good optics
Single focus, white balance, exposure
Full image capture source: researchpedia.info
Eye:
Poor optics
Constantly scanning (saccades)
Constantly adjusting focus
Constantly adapting white balance,
exposure
Mental reconstruction of image (sort of) source: quora.com
Imdad ullah Khan (LUMS) Data Visualization 19 / 101
Visual Perception: Eye vs Camera
Visual Perception is not just camera work
Which square between A and B is darker?
Imdad ullah Khan (LUMS) Data Visualization 20 / 101
Visual Perception: Eye vs Camera
Visual Perception is not just camera work
Both have the same darkness!
Imdad ullah Khan (LUMS) Data Visualization 21 / 101
Visual Perception: Eye vs Camera
Visual Perception is not just camera work
Color is relative
Imdad ullah Khan (LUMS) Data Visualization 22 / 101
Visual Perception: Gestalt Principles
Gestalt Psychology
The human mind considers objects in their entirety before,
or in parallel with, perception of their individual parts;
suggesting the whole is other than the sum of its parts.
Theory of Perception - wikipedia
Gestalt Psychology provides valuable insights into how people perceive
visual components as whole forms rather than just as simple sums of parts
Proximity: Elements close to each other are perceived as a group
Similarity: Items that are similar are grouped together
Continuity: Eyes are drawn along paths, lines, and curves
Closure: We perceive whole shapes even when parts are missing
Anomaly: The mind is very good at identifying outliers
Imdad ullah Khan (LUMS) Data Visualization 23 / 101
Visual Perception: Gestalt Principles
Similarity: The mind perceives similar shapes in a relationship and bring
them together to form larger shapes
How many circles and squares are there?
Imdad ullah Khan (LUMS) Data Visualization 24 / 101
Visual Perception: Gestalt Principles
Anomaly: The mind is very good at identifying outliers
Which piece stands out?
Imdad ullah Khan (LUMS) Data Visualization 25 / 101
Visual Perception: Gestalt Principles
Continuation: The mind finds meaning in continuation in shapes that are
next to each other
Did the leaf come out of the H? Did the “lions” scare the birds?
Imdad ullah Khan (LUMS) Data Visualization 26 / 101
Visual Perception: Gestalt Principles
Closure: The mind makes shapes contiguous
How many triangles? Where is the top of the panda?
Imdad ullah Khan (LUMS) Data Visualization 27 / 101
Visual Perception: Gestalt Principles
Proximity: The mind perceive closer things as related
Is there any big square?
Imdad ullah Khan (LUMS) Data Visualization 28 / 101
Visual Perception: Gestalt Principles
Imdad ullah Khan (LUMS) Data Visualization 29 / 101
Visual Perception: Context
Context: Context can change the appearance of same object
image credit: John Hart, UIUC
Both lines are equal?
Imdad ullah Khan (LUMS) Data Visualization 30 / 101
Visual Perception: Context
Context: Context can change the appearance of same object
image credit: John Hart, UIUC
Which line looks longer?
Imdad ullah Khan (LUMS) Data Visualization 31 / 101
Visual Perception: Context
Context: Context can change the appearance of same object
image credit: John Hart, UIUC
Is the difference more significant?
Imdad ullah Khan (LUMS) Data Visualization 32 / 101
Visual Perception: Context
Context: Context can change the appearance of same object
image credit: John Hart, UIUC
Imdad ullah Khan (LUMS) Data Visualization 33 / 101
Visual Perception: Context
Context: Context can change the appearance of same object
image credit: John Hart, UIUC
Imdad ullah Khan (LUMS) Data Visualization 34 / 101
Visual Perception: Preattention
Preattention: Some visual features are detected immediately
Pop-out vs. Serial Search
If recognition takes 200 − 250ms, then it qualifies as preattentive
eye movements takes > 200ms, yet some processing can be done
quickly
If a decision takes a fixed amount regardless of the number of
distraction, it is considered to be preattentive
It is important for effective visualization to use better discrimination
and avoid misleading viewers
Imdad ullah Khan (LUMS) Data Visualization 35 / 101
Visual Perception: Preattention
Preattention: Some visual features are detected immediately
How many 5’s are there?
385720939823728196837293827
382912358383492730122894839
909020102032893759273091428
938309762965817431869241024
Imdad ullah Khan (LUMS) Data Visualization 36 / 101
Visual Perception: Preattention
Preattention: Some visual features are detected immediately
How many 5’s are there?
385720939823728196837293827
382912358383492730122894839
909020102032893759273091428
938309762965817431869241024
Imdad ullah Khan (LUMS) Data Visualization 37 / 101
Visual Perception: Preattention
Preattention: Color (hue) is preattentive
Detect red circle among these circles
Imdad ullah Khan (LUMS) Data Visualization 38 / 101
Visual Perception: Preattention
Preattention: Form (curvature) is (somewhat) preattentive
Detect red circle among the following objects
Imdad ullah Khan (LUMS) Data Visualization 39 / 101
Visual Perception: Preattention
Preattention: Conjunction of attributes is generally not preattentive
Detect red circle among blue circles and red squares
Imdad ullah Khan (LUMS) Data Visualization 40 / 101
Visual Perception: Preattention
Preattention: Detecting slanted line among vertical lines is preattentive
Imdad ullah Khan (LUMS) Data Visualization 41 / 101
Visual Perception: Selective Attention
Selective Visual Attention: Visual processing confined to certain stimuli
Watch the video of 6 players passing basketballs among themselves
3 players wearing black and 3 wearing white shirts
You should answer with two integers
Counts of the number of aerial and bounced passes between white
shirted players
http://viscog.beckman.uiuc.edu/grafs/demos/15.html
Imdad ullah Khan (LUMS) Data Visualization 42 / 101
Visual Perception: Magnitude Estimation
How much bigger is the bigger circle?
Imdad ullah Khan (LUMS) Data Visualization 43 / 101
Visual Perception: Magnitude Estimation
How much bigger is the bigger bar?
Imdad ullah Khan (LUMS) Data Visualization 44 / 101
Visual Perception: Magnitude Estimation
How much bigger?
Imdad ullah Khan (LUMS) Data Visualization 45 / 101
Visual Perception: Magnitude Estimation
Steven’s Power Law
Heuristics for perceptual estimation
Length is estimated within factors of [.9 − 1.1]
Area is estimated within factors of [.6 − .9]
Volume is estimated within factors of [.5 − .8]
Imdad ullah Khan (LUMS) Data Visualization 46 / 101
Cognitive Load and Information Processing
Effective visualizations reduce cognitive load—making it easier for the
brain to process and understand information.
Intrinsic Load: Complexity inherent in the data itself
Extraneous Load: Complexity added by the way information is
presented
Germane Load: Cognitive effort to process and understand
information
Applying principles of perception and cognition to visualization design
enhances the effectiveness and clarity of visual data representation.
Using color to highlight differences in data points effectively reduces
cognitive load by drawing attention to key elements without overwhelming
the viewer.
Imdad ullah Khan (LUMS) Data Visualization 47 / 101
Basic Principles of Data Visualization
Imdad ullah Khan (LUMS) Data Visualization 48 / 101
Basic Principles of Data Visualization
Understanding the fundamental principles of data visualization is crucial
for creating effective and meaningful visualization
These principles ensure that visualizations are not only appealing but also
functional and informative.
Clarity: The visualization should convey the intended message in a
clear and concise manner
Accuracy: Representations must be precise and accurate to maintain
data integrity
Efficiency: Information should be presented in the most efficient way
possible, without unnecessary complexity
Aesthetics: Visually appealing presentations can engage the
audience more effectively
Imdad ullah Khan (LUMS) Data Visualization 49 / 101
Visualization Principles: Clarity
Clarity is about making the data easy to read and understand. The goal is
to simplify the presentation so that the audience can grasp it quickly
without confusion
Example of Clarity
A bar chart showing sales data over months should have clear labels, a
legible font size, and distinct colors for different products to facilitate easy
understanding.
Imdad ullah Khan (LUMS) Data Visualization 50 / 101
Visualization Principles: Accuracy
Accuracy ensures that the visual representation faithfully reflects the data.
Misleading visuals can lead to incorrect conclusions and decisions
Example of Accuracy
A pie chart representing market share should correctly depict proportions.
Any rounding errors or scaling mismatches can lead to misinterpretation of
the competitive landscape.
Imdad ullah Khan (LUMS) Data Visualization 51 / 101
Visualization Principles: Efficiency
Efficient visualizations convey information quickly and directly, using the
least amount of graphical elements necessary to communicate the message
effectively
Example of Efficiency
A line graph showing trends over time is more efficient than a detailed
table as it allows the viewer to quickly ascertain directional changes and
patterns.
Imdad ullah Khan (LUMS) Data Visualization 52 / 101
Visualization Principles: Efficiency
While functionality is critical, aesthetics play an important role in making
visualizations pleasing to engage with, which can enhance viewer
interaction and retention
Example of Aesthetics
Using a harmonious color scheme and balanced layout in a dashboard can
make the data not only more appealing but also easier to navigate and
interpret.
Imdad ullah Khan (LUMS) Data Visualization 53 / 101
Visual Mapping or Visual Encoding
Imdad ullah Khan (LUMS) Data Visualization 54 / 101
Visual Mapping
Understanding different data types and their appropriate visualization
techniques is essential for effective data representation.
Mapping data attributes to visual attributes
Pick the best mapping
Visually encode different data types for maximum impact and clarity
Consider importance Ordering
Encode the most important information in the most perceptually
accurate way
Imdad ullah Khan (LUMS) Data Visualization 55 / 101
Visual Mapping: Data Types
Imdad ullah Khan (LUMS) Data Visualization 56 / 101
Visual Attributes or Visual Variables
Position
Length
Area
Volume
Shape
Color
Angle
Slope
Texture
Imdad ullah Khan (LUMS) Data Visualization 57 / 101
Relative Magnitude Estimation of Visual Variables
More Accurate
Position
Length
Area
Volume
Shape
Color
Angle
Slope
Texture
Least Accurate
Imdad ullah Khan (LUMS) Data Visualization 58 / 101
Visual Mapping for Data Types
Bertin’s Visual Mapping, Level of Organization
Visual attribute Suitable target data attributes
Position N O Q
Size N O Q
Value N O Q
Texture N O
Color N
Orientation N
Shape N
Imdad ullah Khan (LUMS) Data Visualization 59 / 101
Visual Mapping for Data Types
Mackinlay ranking of attributes by visualization efficacy
Imdad ullah Khan (LUMS) Data Visualization 60 / 101
Visual Mapping: Color Encoding
Imdad ullah Khan (LUMS) Data Visualization 61 / 101
Guidelines for colors
Use only a few colors
Colors should be distinctive and named
Strive for color harmony
Beware of cultural conventions
Beware of bad interactions
Get it right in black and white
Imdad ullah Khan (LUMS) Data Visualization 62 / 101
Evaluating Visualization
Imdad ullah Khan (LUMS) Data Visualization 63 / 101
Evaluating Visualization
Goal of data visualization:
Communicate information clearly and efficiently to users via statistical
graphics, plots, information graphics tables and charts
Effective data visualization is not just about displaying data but doing so
in a way that is accurate, clear, and ethical
These are the criteria to evaluate visualizations
Effectiveness
Expressiveness
Integrity
Consistency
Imdad ullah Khan (LUMS) Data Visualization 64 / 101
Evaluating Visualization: Effectiveness
Effectiveness
A visualization is more effective than another visualization if
the information conveyed by one visualization is more readily
perceived than the information in the other visualization.
Mackinlay, 1986
Keep the design simple and the message clear
Imdad ullah Khan (LUMS) Data Visualization 65 / 101
Effectiveness - Purpose of Visual
Identify purpose of visual - to compare values, show trends, explore
distribution or relationship between variables - choose visual accordingly
https://towardsdatascience.com/tips-for-effective-data-visualization-d4b2af91db37
Imdad ullah Khan (LUMS) Data Visualization 66 / 101
Effectiveness - Focus on Vital Data Points
Vital Data Points are few: Which visual gives better insight of sudden dip?
https://towardsdatascience.com/tips-for-effective-data-visualization-d4b2af91db37
Imdad ullah Khan (LUMS) Data Visualization 67 / 101
Effectiveness - Suppress the Noise
Make the noise less pronounced
https://towardsdatascience.com/tips-for-effective-data-visualization-d4b2af91db37
Imdad ullah Khan (LUMS) Data Visualization 68 / 101
Effectiveness - Use Colors Wisely
Should the same thing be represented with different colors?
https://towardsdatascience.com/tips-for-effective-data-visualization-d4b2af91db37
Imdad ullah Khan (LUMS) Data Visualization 69 / 101
Effectiveness - Avoid Unnecessary Aesthetic Sense
Box Plot with too much aesthetics sense (using too much ink)
Imdad ullah Khan (LUMS) Data Visualization 70 / 101
Effectiveness - Avoid Unnecessary Aesthetic Sense
Scale shifted to side
Imdad ullah Khan (LUMS) Data Visualization 71 / 101
Effectiveness - Avoid Unnecessary Aesthetic Sense
Upper boundaries removed
Imdad ullah Khan (LUMS) Data Visualization 72 / 101
Effectiveness - Avoid Unnecessary Aesthetic Sense
More effective representation
Imdad ullah Khan (LUMS) Data Visualization 73 / 101
Effectiveness - Avoid Unnecessary Aesthetic Sense
Right brightness
Imdad ullah Khan (LUMS) Data Visualization 74 / 101
Effectiveness - Avoid Unnecessary Aesthetic Sense
The following plots have exactly the same information but huge difference
in ink use
https://towardsdatascience.com/tips-for-effective-data-visualization-d4b2af91db37
Imdad ullah Khan (LUMS) Data Visualization 75 / 101
Evaluating Visualization: Expressiveness
Expressiveness
A set of facts is expressible in a visual language if the
sentences (i.e. the visualization) in the language express all
the facts in the set of data and only the facts in the data.
Mackinlay, 1986
Imdad ullah Khan (LUMS) Data Visualization 76 / 101
Evaluating Visualization: Expressiveness
Lengths (interpreted as quantitative values) express non-facts
Imdad ullah Khan (LUMS) Data Visualization 77 / 101
Evaluating Visualization: Expressiveness
Lengths (interpreted as quantitative values) express non-facts
Imdad ullah Khan (LUMS) Data Visualization 78 / 101
Evaluating Visualization: Integrity
Integrity
What is presented should accurately represents what is in
the data being visualized, and that no design choices should
distort or obfuscate the facts and analytical findings
Ensure all visual elements accurately represent the underlying data
Imdad ullah Khan (LUMS) Data Visualization 79 / 101
Evaluating Visualization: Tufte’s Principles of Integrity
1 The representation of numbers, as physically measured on the surface of the
graphic itself, should be directly proportional to the numerical quantities
measured
2 Clear, detailed, and thorough labeling should be used to defeat graphical
distortion and ambiguity. Write out explanations of the data on the graphic
itself. Label important events in the data
3 Show data variation, not design variation
4 In time-series displays of money, deflated and standardized units of
monetary measurement are nearly always better than nominal units
5 The number of information-carrying (variable) dimensions depicted should
not exceed the number of dimensions in the data
6 Graphics must not quote data out of context
Imdad ullah Khan (LUMS) Data Visualization 80 / 101
Evaluating Visualization: Integrity
Students Gender Distribution
https://towardsdatascience.com/tips-for-effective-data-visualization-d4b2af91db37
Imdad ullah Khan (LUMS) Data Visualization 81 / 101
Evaluating Visualization: Integrity
Imdad ullah Khan (LUMS) Data Visualization 82 / 101
Evaluating Visualization: Integrity
Imdad ullah Khan (LUMS) Data Visualization 83 / 101
Evaluating Visualization: Integrity
Distorted x-axis for rise in global warming
Imdad ullah Khan (LUMS) Data Visualization 84 / 101
Evaluating Visualization: Integrity
Yield of a process increased from 56% to 67% over a period of 6 months
Which visual is exaggerating the increase?
Imdad ullah Khan (LUMS) Data Visualization 85 / 101
Evaluating Visualization: Integrity
American election expenditures
Which visual is exaggerating the increase in expenses from 1972 to 1982?
http://www.astro.caltech.edu/ay119/bdass/davidoff-3-viscommfund.pdf
Imdad ullah Khan (LUMS) Data Visualization 86 / 101
Evaluating Visualization: Integrity
Properties of visualization should match the properties of data
Two-dimensional data mapped with three-dimensional representation
Imdad ullah Khan (LUMS) Data Visualization 87 / 101
Evaluating Visualization: Consistency
Consistency: mainly apply to sets of visualizations
Effectiveness and expressiveness ask for optimal visual encoding and space
used in one visual
Individually optimized (locally effective) but not globally consistent visuals
can be misleading
Use consistent styles and colors to avoid confusing the viewer
The same fields should be presented in the same way
Different fields should be presented in different ways
Imdad ullah Khan (LUMS) Data Visualization 88 / 101
Evaluating Visualization: Consistency
Qu & Hullman (2016) Evaluating Visualization Sets:
Trade-offs Between Local Effectiveness and Global Consistency
Imdad ullah Khan (LUMS) Data Visualization 89 / 101
Evaluating Visualization: Consistency
Qu & Hullman (2016) Evaluating Visualization Sets:
Trade-offs Between Local Effectiveness and Global Consistency
Imdad ullah Khan (LUMS) Data Visualization 90 / 101
Evaluating Visualization: Chart Junk
Maximal Data:Ink Rato
A sentence should contain no unnecessary words, a paragraph no
unnecessary sentences, for the same reason that a drawing should
have no unnecessary lines and a machine no unnecessary parts.
William Strunk, Jr.
Do not try to deceive the audience - Avoid manipulating visual
mapping to exaggerate findings
Avoid 3D - visually appealing, but 3D can distort data interpretation
Keep chart junk to minimum to prevent distractions
Minimize use of Ink
Some chart junk helps in remembering though
Excessive use of colors can be distracting and misleading
Imdad ullah Khan (LUMS) Data Visualization 91 / 101
Evaluating Visualization: Chart Junk
Avoid chart junk, if it does not add any value
http://jcsites.juniata.edu/faculty/rhodes/ida/graphicalIntRedes.html
Imdad ullah Khan (LUMS) Data Visualization 92 / 101
Evaluating Visualization: Chart Junk
Avoid chart junk, if it does not add any value
Imdad ullah Khan (LUMS) Data Visualization 93 / 101
Evaluating Visualization: Chart Junk
Avoid chart junk, if it does not add any value
Imdad ullah Khan (LUMS) Data Visualization 94 / 101
The Grid System
Grid system naturally organizes data to give it more meaning
Imdad ullah Khan (LUMS) Data Visualization 95 / 101
The Grid System
Which news is more important? Which is more visible?
Imdad ullah Khan (LUMS) Data Visualization 96 / 101
The Grid System
Grouping of elements in columns has a certain meaning
Imdad ullah Khan (LUMS) Data Visualization 97 / 101
Data Visualization Process
From data to insight, the visualization process involves several key steps:
1 Data Collection: Gathering the necessary data from various sources
2 Data Cleaning: Preparing the data by cleaning and structuring it
3 Data Analysis: Analyzing the data to find patterns and insights
4 Data Visualization: Representing the data visually to highlight
findings
5 Insight Communication: Using the visualization to tell a story or
support decision-making
Imdad ullah Khan (LUMS) Data Visualization 98 / 101
Data Visualization Process: Purpose
Purpose of your visualization
Are you exploring the data?
Are your formatting it for decision making?
Or are you telling a story?
Imdad ullah Khan (LUMS) Data Visualization 99 / 101
Data Visualization Process: Principles
Eight Principles of communicating through data
Define what questions are you answering
Use accurate data
Experiment with ways to answer
Go with cognitive research (go with the rules defined through
previous research for data visualization)
Faithfully represent your data
Tailor it to your audience
Make it as simple as possible
Remove everything that you can
Imdad ullah Khan (LUMS) Data Visualization 100 / 101
Data Visualization Process
1 Choosing the visualization for your purpose
Simple numbers? pie charts? bar charts? Tables? plots? maps?
2 Choosing right tool and coding language
Excel, tableau, Microsoft power BI, illustration software
R, Python etc.
Imdad ullah Khan (LUMS) Data Visualization 101 / 101
