Why Create Visualizations?
Answer questions (or discover them) Make decisions/See data in context/Expand memory/Find patterns or discover errors in the data
Communicate information to others/Share and persuade/Collaborate/Inspire. PC based visualization systems provide visual representations of datasets designed to help
Attribute types
Types
peopleofcarry
dataout tasks more effectively. Categorical (nominal)
Charts can be selected based on the types of dimensions and measures. At a high level, data attributes can be divided into 3 -compare equality
principal types (continuous, ordinal, and categorical). Temporal and geographical data are often special cases that have -no implicit ordering
specific mappings to chart types. Ordered
Continuous: ordered, equally and meaningfully spaced values -ordinal
Ordinal: discrete values that can be ordered e.g. by ranking Less/greater
defined than
Categorical: discrete values that have no particular ordering
Temporal data: interval data with a time component -quantitative
Geographical data: inherently two-dimensional data that refers to places Meaningful magnitude
Arithmetic possible
Cognitive load is the effort or brainpower that it takes to process something. In this case the visual elements present in the charts and dashboards you design. In general your aim is to reduce
unnecessary cognitive load by eliminating the need to process elements that take up mental resources but do not help your audience understand the insights or information you are sharing. To
achieve this we want to boost the signal, i.e. more effectively communicate information, and reduce the noise, i.e. remove clutter or visual elements that take up space but don’t increase
understanding.
Perception and cognition: Seeing is fast and efficient, using preattentive processing. Conscious thinking takes more time and effort.
Gestalt psychology: Understanding how humans perceive pattern, form, and organization, and how this can be used in design.
Visual variables: Understanding principles behind ways in which graphic symbols can be used to represent and distinguish between data items. In most business context, it is best to use a
sans=serif front. They have better readability in screens and carry the perception of being contemporary and modern. Some sans-serif examples are below Helvetica Graphik Calibri Aria.
Some ways of representing Temporal Data:
Data ink ratio= 'ink' uses that directly encodes that data/total ‘ink' used in the viz. Line charts: show how measures changes over continuous time. Bar chart: Show measures within discrete time units
Aim for minimalism in your: Backgrounds; Axes; Borders; Gridlines; Colors; Images; Labeling. Heatmap: show variation within a range over time. Slope charts: show how data between two points before and after
Be wary of chartjunk14 – the elements of a viz that do not directly explain the
like pictures, icons, or additional embellishment can be confusing. Temporal/Time data data. Things
(Type: continuous)is simply information that relates to time and is very commonly seen in
charts. Slope chart: Aim is to show a comparison between two periods of time, E.g. ranking Mereator projection: Shapes are preserved, areas are severely distorted. Goode homolosing: Areas are preserved.
in two different years or sales comparison between two quarters.
Two time dimensions: Heat map (show seasonality, Calendar year in a table Geospatial data has location information that can be portrayed on a map. Map visualizations commonly used in socioeconomic and political
Cycle plot: A cycle plot needs to be quite wide. contexts; e.g., visualizing election results
Alternative — show only the times you are interested in
Alternative — show only the times you are interested in Key reasons for using a map include:
Alternative - Line chart alternative to a cycle plot that uses less width.
Location-based storytelling. Is location central to the insight you want to share? E.g. highlighting localized outbreaks of disease or showing which
branches of a store are most profitable.
Spatial relationships. Are there patterns or trends that are best understood by viewing the location of data points relative to each other? E.g.
visualizing origin and destination points for transport routes, or house prices relative to schools or subway stations.
Geographic overview. Is the geography familiar enough for a data overview to be more quickly and easily digested in a map? E.g. showing
election results.
