Data Visualization
Data Visualization - Overview
“the strength of data visualizations come from their
appeal and their usability for readers. You often only
have 5-to-10 minutes to capture the attention of a
policymaker; therefore the visualization needs to have
a strong focus” (IDRC Grantee Respondent, 2012).
• Important for understanding and gain insights from data.
• Organizations of all types and sizes are generating data each minute, hour and day.
• Learn things from data that can help make better decisions, take smarter actions and operate
more efficiently. 2
Exercise: Past, Present and Future World
Development
• How has the world developed in last 200 years?
• Consider wealth/poverty and health as indicators of human
condition.
• Measurable indicators: lifespan and income per capita.
• Other interesting variables: region, country, population.
• Lets find data
GDP per capita 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
Abkhazia
Afghanistan 603 603 603 603 603 603 603 603 603 603 604
Akrotiri and Dhekelia
Ex 1: Past, present and future World development
Albania
Algeria
American Samoa
667
716
667
716
Multiple Sources
668
717
668
718
1. Human Mortality
668
719
668
720
668
721
668
722
668
723
668
724
668
725
Andorra 1197 1199 1201 1204 1206 1208 1210 1212 1215 1217 1219
Angola 618 620 623 626 628 631 634 637 640 642 645 Database,
Anguilla
[Link]
• How has the world developed in the past 200 years?
Antigua and Barbuda 757 757 757 757 757 757 757 758 758 758 758
Argentina 1507 1508 1508 1508 1508 1508 1509 1509 1509 1509 1510
Armenia 514 514 514 514 514 514 514 514 514 514 514
Aruba 833 833 833 833 833 833 833 834 834 834 834
• Consider Wealth/Poverty and Health as Indicators of the 2.
Australia
Austria
815
1848
816
1855
818
1863
820 World Population
Human 822
Prospects: The 2010
1870 1878
824
1885
826
1893
828
1901
830
1908
832
1916
834
1924
Condition
Azerbaijan
Bahamas
775
1445
775
1445
775
1445
775 775
Revision / United
1446 1446
776
1446
776
1446
776
1447
776
1447
776
1447
777
1447
• Measurable Indicators: Lifespan and Income per Capita Nations
Life expectancy
Abkhazia
1800 1801 1802 1803Population
Division, with
1804 1805 1806 1807 1808 1809
Afghanistan 28.21 28.2 28.19 28.18 28.17 28.16 28.15 28.14 28.13 28.12
• Other interesting variables: Region, Country, Population projections
Akrotiri and Dhekelia
Albania
Algeria
35.4
28.82
35.4
28.82
35.4
28.82
35.4
28.82
35.4
28.82
35.4
28.82
35.4
28.82
35.4
28.82
35.4
28.82
35.4
28.82
3. Global Burden of
• Let’s go find Data:
American Samoa
Andorra
Angola 26.98 26.98 26.98
Disease Study 2015
26.98 26.98 26.98 26.98 26.98 26.98 26.98
Anguilla
Antigua and Barbuda 33.54 33.54 33.54 33.54 33.54 33.54 33.54 33.54 33.54 33.54
Argentina
Armenia
33.2
34
33.2
34
33.2
34
33.2
34
33.2
34
33.2
34
33.2
34
33.2
34
33.2
34
33.2
34
> 200 countries, >
Aruba 34.42 34.42 34.42 34.42 34.42 34.42 34.42 34.42 34.42 34.42 200 years, ~120k
Australia 34.05 34.05 34.05 34.05 34.05 34.05 34.05 34.05 34.05 34.05
Austria 34.4 34.4 34.4 34.4 34.4 34.4 34.4 34.4 34.4 34.4 data points
Azerbaijan 29.17 29.17 29.17 29.17 29.17 29.17 29.17 29.17 29.17 29.17
Bahamas 35.18 35.18 35.18 35.18 35.18 35.18 35.18 35.18 35.18 35.18
Bahrain 30.3 30.3 30.3 30.3 30.3 30.3 30.3 30.3 30.3 30.3
Bangladesh 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5
Barbados 32.12 32.12 32.12 32.12 32.12 32.12 32.12 32.12 32.12 32.12
Video Information
• Hans Rosling (27 July 1948 – 7 February 2017) was a Swedish
physician, academic, statistician, and public speaker. He was the
Professor of International Health at Karolinska Institute and was the
co-founder and chairman of the Gapminder Foundation, which
developed the Trendalyzer software system. He held presentations
around the world, in which he promoted the use of data to explore
development issues.
What Data Visualization techniques?
Variable Type Visual Representation
Life Expectancy Quantitative (discrete) Position (y)
Income Quantitative (continuous) Position (x)
Total Population Quantitative (discrete) Size
Geographical Region Nominal Colour
Time Ordinal Animation
• Artistic Principles
• Incrementalism
• Expressiveness / Storytelling
• Interactivity connects directly to the Data (A Conversation)
• Exploratory vs Explanatory
Ex.2 Minard’s Classic Graphical Representation of Napoleon’s
Demise
What Data Visualization techniques?
Variable Type Visual Representation
Size of the army Quantitative (discrete) Size
Location Quantitative (continuous) Position (x,y) – spatial map
Direction of Army Nominal Colour
movement
Temperature during Quantitative (continuous Line graph
Retreat
Time Ordinal Line graph
• Artistic Principles
• Expressiveness / Storytelling
• Minard’s intuitive sense for visceral effect
• According to Edward Tufte – “may well be the best statistical graphic ever drawn”
• Exploratory vs Explanatory
How to manage Data?
• With so much data the challenge is how to display results of data
exploration and analysis in a way that is not overwhelming.
• Need a new way to look at the data – collapses and condenses the results
in an intuitive fashion but still display graphs and charts that decision
makers are accustomed to.
• In today’s on-the-go society, results needed quickly via mobile devices, and
capability to easily explore data on own in real time should be provided.
• Data size and column composition play an important role when selecting
graphs to represent your data.
10
KPI
Visualization Process Model
Process Actions
• Business Context
• What Questions?, Audience Business
Consumers
User
• Get Data
Question Experience
• Get your data
• Find it, clean it, prepare it Decisions Business
• Visualization design consideration & Actions Context
• User experience
• How will the visualization be used?
• How should the data be encoded?
• Build visualizations Get
• Ask questions of the data Business insights
data
• Build structure & visualize data
• Explore data and focus analysis Pie
• Encoding data to a visual object Map
Visualization Design
• Business Insights Bar
Build considerations
• Generate Graphics, Charts
• Derive Insights Visualizations STRATEGIC
• Decisions & Actions ANALYTICAL
• Share insights with others in views,
dashboards, and stories OPERATIONAL
Business Context - Visualization Goals
• Answer Questions
• Make decisions
• See data in context
• Support graphical calculations
• Find patterns
• Present argument or tell a story
• Inspire
Business Context – Business Question
• The business questions should focus on the following aspects of the
organization:
• What – e.g. what are the products that are selling the most?
• When – e.g. at what times of the year do the sales of a certain product
spike?
• Where – e.g. what regions/countries is a particular product doing best in?
• As compared to – e.g. how have the sales of Brand X’s 50ml water sold
compared to Brand Y’s over the last 3 months?
• By how much – e.g. what is the percentage increase in the sales of Product
X over the last 3 months?
Design Considerations
Types of Users – levels of decision making
The narrative will differ depending on the use of the visualization.
• Strategic- monitor the execution of corporate strategic objectives at each level
of the organization and emphasize management more than monitoring or
analysis. Strategic dashboards are often implemented using the Balanced
Scorecard methodology and are referred to as scorecards.
• Analytic- track departmental processes and projects and emphasize analysis
more than monitoring or management. They are often implemented using
portals and run against data marts or warehouses where data is refreshed
periodically.
• Operational – track core operational processes and often display real-time
data. Operational dashboards emphasize monitoring more than analysis or
management.
Data Visualization is about
Creating a Narrative…
User Experience on 3 Levels
• Visceral – relates to the quick emotional reaction about good and bad. This feeling is
more intuitive than rational and involves the basic perceptual tasks of distinguishing
objects and is really about forming the true first impression.
• Behavioral – this experience is about conscious awareness and is controlled by
cognition but influenced by visceral reaction. The behavioral experience focuses on
issues such as readability and usability. These visualizations are interactive in nature
and allow users to literally have a conversation with the data.
• Reflective – this experience is solely a cognitive and intellectual response. It is not
influenced by the visceral reaction. Reflective user experience occurs in
visualizations where the graphics results in decisions and/or actions.
Data Visualization is about
Creating a Narrative…
Data Visualization is about Creating a Narrative…
Data Visualization is about
Creating a Narrative…
Encoding data
• The various dimensions of data can be encoded with appropriate visual
properties. There are a number of options for data encoding including color,
size, shape, spatial position, length, width and orientation and the choice of
these is extremely important for the quality of the visual
Categorical Data Ordinal Data Quantitative Data
Position Position Position
Colour Hues Size Length
Shape Colour Intensity Size
Clusters Colour Hues Colour Intensity
Boundaries Shapes
Does Size matter?
Does Colour matter?
Build Visualizations
• In building visualizations, the data needs to be encoded as a visual object.
• There are multiple components that need to be part of an visualization.
• For example, if the focus of your organization is on customer satisfaction then
you need to graphically encode
• the current customer satisfaction levels
• customer satisfaction across branches/regions
• trend of customer satisfaction across years, quarters
• relationship between sales and customer satisfaction etc.
• For each of the graphs the analyst will need to determine the most appropriate
(i.e. bar, pie, line, heat map), connect them so that the big picture emerges.
Business Insights
• The visualizations enable the decision makers to gain evidence-based insights
into the organization.
• The business user interacts with the visualizations to derive insights based on
the compelling stories that emerge from various graphics such as the charts,
maps and graphs collectively.
• The graphical, summarized and multiple views of data enable the decision
makers to understand the complete picture that is embedded in it.
Decisions and Actions
• This step involves sharing insights with others through views, dashboards, and
stories.
• These insights are then used to drive business actions and make more data
driven decisions.
• The impact of the visualization can be realized when the decisions and/or
actions improve the performance of the organization.
• The visualizations may explicate some of the mysteries embedded in the data
which will guide new business questions.
Summary –
How to generate the best visuals for data?
• Determine what you are trying to visualize and what kind of
information you want to communicate.
• Know your audience and understand how it processes visual
information.
• Understand the data you are trying to visualize, including its size
and cardinality
• Use a visual that conveys the information in the best and simplest
form for your audience.
27
Build Visualizations - Techniques
28
Tables (Cross Tabs/Pivot Tables)
29
Highlight Tables
30
31
Line Graphs/Chart
• Shows the relationship of one variable to another.
• Most often used to track changes or trends over time.
• Useful when comparing multiple items over the same time
period
• The number of data points available may dictate the best
visual to use
• Excellent for tracking multiple data sets on the same chart to
see any correlation trends.
[Link]
Track Changes Over Time
Line Graphs/Chart Relationship between variables
X axis – time (continuous)
Y axis - measure
33
Bar Charts
• Most commonly used for comparing the quantities of different
categories or groups.
• Configured with either vertical or horizontal bars, with the length or
height of each bar representing the value.
• Suited when values are distinct enough that differences in the bars
can be detected by the human eye.
• When the values (bars) are very close together or there are large numbers of
values (bars) more difficult to compare the bars to each other.
34
Comparing the quantities of different categories or groups. 35
Bar Chart: ChickV
36
37
Pie Charts
• Used to compare the parts of a whole.
• Pie Charts can be difficult to interpret because the human eye has a
hard time estimating areas and comparing visual angles.
• Another challenge is that it is difficult to compare slices of the pie
that are similar in size but not located next to each other.
• Most effective when there are limited components and when text and
percentages are included to describe the content.
• When creating dashboards another problem is the amount of space
required.
38
Pie Chart
39
Bad Pie Charts
40
Scatter Plots
• Helps gain a sense of how spread out the data might be or how
closely related the data points are.
• Can quickly identify patterns present in the distribution of the data.
• Used to visualize relationships between numerical variables.
• Scatter plots are helpful when there are many data points.
• If small set of data points, a bar chart or table may be more effective.
41
Scatterplot
42
Scatter Plot: department/Parish/Sales/Profit
43
Bubble Plots
• A variation of a scatter plot.
• Shows the relationship among at least 3 measures
• Two measures represented by the plot axes.
• Third measure represented by the size of the bubbles.
• Useful for data sets with dozens to hundreds of values or
• When the values differ by several orders of magnitude
• You can use color to represent an additional measure, and you can
animate the bubbles to display changes in the data over time.
• A geo bubble map is a bubble plot that is overlaid on a geographic
map.
44
Bubble Plot
Sum(Profit)
Profit
Sales 45
46
Heat Maps
• Used to compare categorical data using color.
• Individual values in a matrix are represented as colors.
• Size and/or color emphasize points
47
Heat Maps
48
Heat Maps
49
Symbol Maps
• Using maps is a very powerful tool that can quickly show the user
geographical data at a glance.
• Informs users about their data in relation to location.
• Must have longitude or latitude points or have fields of data that can
be matched to geographical roles (e.g. country, city).
50
51
Symbol Map: Chick-V Cases in Jamaica
52
53
54
Filled Map
55
Filled Map
Geocoding is used to fill a
polygon with a color
Sum of profit by state
56
Treemaps
• Display data in nested rectangles.
• Dimensions define the structure of the treemap and measures to
define the size or color of the individual rectangles.
• Ideal for hierarchical and part-to-whole relationships.
57
Sales
Profit
Sales = Size
Color = Profit
58
59
60
Showing Data vs
Storytelling With Data
“There is a story in your data. But your tools don’t know what that story is. That’s where it
takes you— the analyst or communicator of the information— to bring that story visually
and contextually to life.”
Nussbaumer Knaflic, Cole. Storytelling with Data: A Data Visualization Guide for Business Professionals (Kindle Locations 478-480). Wiley.
Kindle Edition.
61
62
63
Showing Data
64
Storytelling With Data
65
Showing Data
66
Storytelling With Data
67
Conclusion: Data Visualization Tips
• Start with the questions.
• Know what you are trying to say.
• Visualization must have a purpose and must select what to include to fulfill
that purpose.
• Once purpose is clear need to decide which analysis supports that purpose
e.g. Trends over time
• The selection of charts usually will depend upon the number of categories
and measures (or dimensions) you want to visualize.
• The audience and purpose of the visualization should always be top of
mind when considering what you want to visualize.
68
Making your visualization more effective
• The single most important step to make a great visualization is to
know what you’re trying to say.
• How do you know if your visualization has a purpose? Ask some
questions to find out.
• Who is your audience?
• What questions do they have?
• What answers do you find for them?
• What other questions does it inspire?
• What conversations will result?
69