This is a presentation I gave on Data Visualization at a General Assembly event in Singapore, on January 22, 2016. The presso provides a brief history of dataviz as well as examples of common chart and visualization formatting mistakes that you should never make.
3. “You can have piles of facts
and still fail to resonate. It’s
not the information itself
that’s important but the
emotional impact of that
information.”
“[few] grasp how to use data
to tell a meaningful story that
resonates both intellectually
and emotionally with an
audience”
Nancy Duarte – Writer, Speaker, CEO Daniel Waisberg - Analytics Advocate,
Google
5. 5
AN INTRODUCTION
4.3
2.5
3.5
4.5
2.4
4.4
1.8
2.8
2 2
3
5
0
2
4
6
Singapore Hong Kong Indonesia Philippines
Volume
2013 2014 2015
58%23%
10%
9%
1st Qtr
2nd Qtr
3rd Qtr
4th Qtr
0
2
4
6
8
Category
1
Category
2
Category
3
Category
4
Category
5
Category
6
Category
7
Category
8
Category
9
Category
10
Volume
Series 1 Series 2 Series 3
0
10
20
30
40
0 5 10 15 20 25
Time
Occurrence
but data visualization starts with the basics
6. 6
AN INTRODUCTION
a primary goal of data visualization is to communicate information clearly and
efficiently to users via the statistical graphics, plots, information graphics, tables,
and charts selected
data visualization
the visual representation of data
“the purpose of visualization is insight, not pictures”
- Ben Shneiderman, computer scientist
13. 13
Although this wasn’t the first use of the pie chart, Minard popularized it’s use with his
graphic depiction of cattle volumes and distribution throughout France. Remember, this
was hand drawn!
15. 15
A BayArea (Redwood CityWharf) tide prediction diagram for each 24-hour day in the
Month of June 2013
Source: http://www.informationisbeautifulawards.com/showcase/150-tide-predictions
16. 16
Thematic map that shows the distribution and biodiversity of NewYork City’s street trees
based on the last tree census.
Source: http://jillhubley.com/project/nyctrees/
18. 18
what’s wrong with it?
No Y axis
Multi-axis not declared and labelled
what is it?
Chart shown by Rep Jason Chaffetz
during Planned Parenthood hearing
Watch it here - http://bit.ly/1UhTYoU
X-axis spans 8 years, but chart only
includes 2 years (2006 and 2013)
Example 1 – Lying with data
19. 19
Example 1 – Lying with data
Planned Parenthood – Abortions vs Cancer Screening Services
Bar Chart | Single Axis | Vertical
how to fix it?
if you are comparing data between 2 non-
adjacent years a bar chart would be a
better choice
here’s the same data in a bar chart, and
on a single axis
20. 20
Example 1 – Lying with data
Planned Parenthood – Abortions vs Cancer Screening Services
how to fix it?
Line Chart | Single Axis
Better yet, if you retrieve the data for the
missing years you can use a line chart to
show the change over time
here’s the same data in a line chart with
the missing years (2007 – 2012) included.
*Note, 2008 is missing as I couldn’t find the report for that year
21. 21
Example 1 – Lying with data
Planned Parenthood – Abortions vs Cancer Screening Services
how to fix it?
And for good measure, here’s the same
data represented using a multi axis chart,
and with the axis’ properly labelled
Line Chart | Multi Axis
22. 22
Example 1 – Lying with data
Stacked Bar Chart |
Horizontal
Here’s a slightly different way to look at this data. The above is a stacked bar chart, which shows the % share of all services
PP offers, not just Abortions and Cancer Screening. Notice that, as a % of total services, Abortions is constant at 3% every
year. Read more about this chart on my blog - http://bit.ly/1nrbi0x
how to fix it?
23. 23
Example 2 – Lying with data
Line Chart | Single Axis
what’s wrong with it?
Y axis inflated, removes reader’s ability to
see meaningful change in data
what is it?
national review (@NRO) chart depicting
change in global average temperature
over time
See it here - http://bit.ly/23klTus
Chart does not convey context. The
increase in global average temp of just 2
degrees can have significant effects on
our planet, so using 120 point scale
removes important context and makes
this chart impossible to read
Average Global Temperature (Fahrenheit)
24. 24
Example 2 – Lying with data
Figure 5. Average Global Temperature (Fahrenheit) - relative scale
Line Chart | Single Axis
how to fix it?
Adjust the y-axis so it shows the full
context of the data.
25. 25
what’s wrong with it?
Too much non-data ink
Poor chart labeling
what is it?
Chart shown by Fox News that shows
Average Annual economic growth in the
USA.
Unequal time intervals
Example 3 – Deceptive
Visualization
26. 26
Example 3 – Deceptive
Visualization
how to fix it?
Strip away all unnecessary formatting (e.g.
data ink)
See more here - http://bit.ly/2eIpC18
30. 30
Example 4 – The parts don’t add up to a whole
Pie Chart
what’s wrong with it
charted values different from actual
values
data labels in legend instead of on chart
what is it?
Infographic about U.S. lifestyle trends
31. 31
Example 4 – The parts don’t add up to a whole
Pie Chart
what’s wrong with it
charted values different from actual
values
data labels in legend instead of on chart
what is it?
Infographic about U.S. lifestyle trends
32. 32
Example 4 – The parts don’t add up to a whole
Pie Chart
how to fix it?
The reason this happened was because
the author didn’t have a firm grasp of the
dataset and how it was structured. For
example, a pie chart would actually work
here if the data series was inverted so
each chart represents a year and is broken
down by the desire states, instead of the
other way around.
33. 33
Example 5 – Wrong chart
Tree Map
what is it?
what’s wrong with it
infographic about millennials and brunch
box proportions aren’t accurate
wrong choice of chart, data isn’t
hierarchical
box proportions aren’t accurate
34. 34
Example 5 – Wrong chart
Tree Map
how to fix it?
Always ensure you choose the right chart
that enables you to accurately and
meaningfully present your data. Here’s a
hand resource which helps you select the
right chart based on what you’re
attempting to show (i.e. distribution,
composition, comparison, etc).
Also, when constructing your charts don’t
ever eyeball scale, proportion or distance.
If you create charts manually by hand you
run the risk of misrepresenting the data
35. 35
Example 6 – Confusing chart
Radial Bar Chart
what is it?
what’s wrong with it
Infographic on JD.com sales and growth
missing important visual cues
rings shouldn’t go 360 degrees
36. 36
Example 6 – Confusing chart
Radial Bar Chart
how to fix it?
Make sure you chart is intuitive and easy
to read, and never sacrifice readability for
creative flare. For radial bar charts follow
the basic formatting rules; don’t design
the chart so the rings go a full 360
degrees and clearly label your data.
37. 37
Example 7 – Charts within charts
Pie Chart
what’s wrong with it
how to fix it?
charts within charts
convoluted datapoints and interpretation
don’t embed charts within charts
don’t make data needlessly confusing. If
it’s difficult to interpret your not finished
Read more about how to fix this chart:
http://bit.ly/1nxbZ8t
38. 38
Example 8 – Meaningless visualization
what’s wrong with it
how to fix it?
bubble size has no relation to value
inside
unintentional Venn diagram gives data
new meaning
Always think about how your design
choices impact the interpretation of your
data and design with purpose
Read more about how to fix this chart:
http://bit.ly/1nxbZ8t
39. 39
Example 8 – Meaningless visualization
what’s wrong with it
Too many variables on a single chart (e.g.
year, response state, % of respondents)
46. 46
good
always include axis
titles
add figure labels and use a descriptive
title
always include a legend
Use colours effectively
start with zero axis baseline and only adjust if it helps to provide useful context, and so long as it’s
truthful
48. 48
conduct a readability test
Find some people to look at your chart or infographic
Let them look at it for 10 seconds
Ask them 2 questions
“What was the underlying dataset about?”
“What conclusion or conclusion(s) did you take away from it?”
1.
2.
3.
If anyone didn’t understand what the data was about, you need to rethink your design. If
their conclusions are not aligned with your intended message or their conclusions are wildly
different across your test audience, you’ve got more work to do.
53. 53
Don’t just fill a
dashboard or report
with meaningless
charts. Every chart
and graphic should
have a specific
purpose… and should
be formatted
correctly
61. 61
LESSONS
some resources
Flowing Data – Basic Rules for Making Charts - http://bit.ly/1nqd3v7
Visage: Data Viz 101 - http://bit.ly/1ncO4dK
National Geographic – Spotting Charts that Lie - http://bit.ly/1Guozac
Analythical - My Blog! – analythical.com
Sources:
Left by Pedro Monteiro: https://whatype.wordpress.com/2008/10/12/what-type-of-events-affects-oil-prices/
Right by Kir Khachaturov: http://www.informationisbeautifulawards.com/showcase/113-arab-spring
Sources:
Left by Pedro Monteiro: https://whatype.wordpress.com/2008/10/12/what-type-of-events-affects-oil-prices/
Right by Kir Khachaturov: http://www.informationisbeautifulawards.com/showcase/113-arab-spring
Source:
http://www.informationisbeautifulawards.com/showcase/150-tide-predictions
Full - http://iibawards-prod.s3.amazonaws.com/projects/images/000/000/150/large.jpg?1403718826