Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Speaker Deck
PRO
Sign in
Sign up for free
Telling stories with data visualization
Chris Keathley
November 08, 2017
Programming
0
290
Telling stories with data visualization
A discussion about best practices in information design. Presented at oredev.org
Chris Keathley
November 08, 2017
Tweet
Share
More Decks by Chris Keathley
See All by Chris Keathley
Building Adaptive Systems
keathley
25
1.1k
Contracts for building reliable systems
keathley
3
400
Kafka, the hard parts
keathley
1
900
Building Resilient Elixir Systems
keathley
5
1.3k
Consistent, Distributed Elixir
keathley
4
990
Easing into continuous deployment
keathley
1
110
Leveling up your git skills
keathley
0
420
Generative Testing in Elixir
keathley
0
210
Concurrent Feature tests with Wallaby
keathley
2
410
Other Decks in Programming
See All in Programming
A Philosophy of Software Design 後半
yosuke_furukawa
PRO
10
2.9k
Get Ready for Jakarta EE 10
ivargrimstad
0
1.2k
The strategies behind ddd – AdeoDevSummit 2022
lilobase
PRO
5
260
Let's keep Commodore 64 alive for the next 40 years
mehowte
1
110
Why Airflow? & What's new in Airflow 2.3?
kaxil
0
120
データ分析やAIの "運用" について考える
mmorito
0
150
無限スクロールビューライブラリ 二つの設計思想比較
harumak
0
270
Jetpack Composeでの画面遷移
iwata_n
0
200
Java アプリとAWS の良い関係 - AWS でJava アプリを実行する一番簡単な方法教えます / AWS for Javarista
kanamasa
2
1.3k
Reactは何を提供するLibraryなのか?
taro28
5
630
CSE360 Tutorial 07
javiergs
PRO
0
100
LINE Messaging APIの概要 - LINE API総復習シリーズ
uezo
1
190
Featured
See All Featured
The MySQL Ecosystem @ GitHub 2015
samlambert
238
11k
Teambox: Starting and Learning
jrom
123
7.7k
Why Our Code Smells
bkeepers
PRO
324
55k
GraphQLの誤解/rethinking-graphql
sonatard
28
6.6k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
7
1.1k
Facilitating Awesome Meetings
lara
29
4k
GitHub's CSS Performance
jonrohan
1020
420k
Designing Experiences People Love
moore
130
22k
The Invisible Customer
myddelton
110
11k
We Have a Design System, Now What?
morganepeng
35
3k
Building an army of robots
kneath
299
40k
StorybookのUI Testing Handbookを読んだ
zakiyama
5
2.3k
Transcript
TELLING STORIES WITH DATA VISUALIZATION Chris Keathley / @ChrisKeathley /
c@keathley.io
None
Lets talk about…
Lets talk about… Visualizations
Lets talk about… Visualizations Common mistakes and how to fix
them
Lets talk about… Visualizations Common mistakes and how to fix
them Case Study
Visualizations are important
“Numerical calculations are exact, but graphs are rough” - other
people
55.3846,97.1795 51.5385,96.0256 46.1538,94.4872 42.8205,91.4103 40.7692,88.3333 38.7179,84.8718 35.641,79.8718 33.0769,77.5641 28.9744,74.4872 26.1538,71.4103
23.0769,66.4103 22.3077,61.7949 22.3077,57.1795 23.3333,52.9487 25.8974,51.0256 29.4872,51.0256 32.8205,51.0256 35.3846,51.4103 40.2564,51.4103 44.1026,52.9487 46.6667,54.1026 50,55.2564 53.0769,55.641 56.6667,56.0256 59.2308,57.9487 61.2821,62.1795 61.5385,66.4103 61.7949,69.1026 57.4359,55.2564 54.8718,49.8718 52.5641,46.0256 48.2051,38.3333 49.4872,42.1795 51.0256,44.1026 45.3846,36.4103 42.8205,32.5641 38.7179,31.4103 35.1282,30.2564 32.5641,32.1795 30,36.7949 33.5897,41.4103 36.6667,45.641 38.2051,49.1026 29.7436,36.0256 29.7436,32.1795 30,29.1026 32.0513,26.7949 35.8974,25.2564 41.0256,25.2564 44.1026,25.641 47.1795,28.718 49.4872,31.4103 51.5385,34.8718 53.5897,37.5641 55.1282,40.641 56.6667,42.1795 59.2308,44.4872 62.3077,46.0256 64.8718,46.7949 67.9487,47.9487 70.5128,53.718 71.5385,60.641 71.5385,64.4872 46.9231,79.8718 48.2051,84.1026 50,85.2564 53.0769,85.2564 55.3846,86.0256 56.6667,86.0256 56.1538,82.9487 53.8462,80.641 51.2821,78.718 50,78.718 47.9487,77.5641 29.7436,59.8718 29.7436,62.1795 31.2821,62.5641 57.9487,99.4872 61.7949,99.1026 64.8718,97.5641 68.4615,94.1026 70.7692,91.0256 72.0513,86.4103 73.8462,83.3333 76.6667,75.2564 77.6923,71.4103 79.7436,66.7949 81.7949,60.2564 83.3333,55.2564 85.1282,51.4103 86.4103,47.5641 87.9487,46.0256 89.4872,42.5641 93.3333,39.8718 95.3846,36.7949 98.2051,33.718 56.6667,40.641 59.2308,38.3333 60.7692,33.718 63.0769,29.1026 64.1026,25.2564 64.359,24.1026 74.359,22.9487 71.2821,22.9487 67.9487,22.1795 76.6667,75.2564 77.6923,71.4103 79.7436,66.7949 81.7949,60.2564 83.3333,55.2564 85.1282,51.4103 86.4103,47.5641 87.9487,46.0256 89.4872,42.5641 93.3333,39.8718 95.3846,36.7949 98.2051,33.718 56.6667,40.641 59.2308,38.3333 60.7692,33.718 63.0769,29.1026 64.1026,25.2564 64.359,24.1026 74.359,22.9487 71.2821,22.9487 67.9487,22.1795 65.8974,20.2564 63.0769,19.1026 61.2821,19.1026 58.7179,18.3333 55.1282,18.3333 52.3077,18.3333 49.7436,17.5641 47.4359,16.0256 44.8718,13.718 48.7179,14.8718 51.2821,14.8718 54.1026,14.8718 56.1538,14.1026 52.0513,12.5641 48.7179,11.0256 47.1795,9.8718 46.1538,6.0256 50.5128,9.4872 53.8462,10.2564 57.4359,10.2564 60,10.641
None
None
Alberto Cairo @AlbertoCairo thefunctionalart.com
Anscombe’s Quartet
None
None
None
Tons of tools
D3 graphviz matplotlib R Canvas Emacs org mode
“telling stories”
“The viewer can see the entire narrative” - me just
now
None
I live here
“Some people are less than a mile from [the utility’s]
service area but, by law, can't get its broadband”
None
“Facebook’s offer price was $38 a share, giving the company
a valuation of $104 billion, nearly four times larger than google in 2004” - Washington Post
None
None
None
None
None
None
http responses
None
FICTION
None
Unicorn
The real world Unicorn
None
VISUALIZATION IS ONE VIEWPOINT
“THIS IS HOW THE SERVER IS DOING”
“THIS IS HOW OUR USERS ARE DOING”
None
(UNCONCIOUS) BIAS
Common mistakes and how to solve them
Pie Charts
None
Don’t use pie charts
Colors
Color is filled with subtle bias
None
Use color deliberately and sparingly
None
Avoid using red and green in the same display
ChartJunk and Noise
Initial design goals
Initial design goals
small multiples
Initial design goals
Sparklines
Initial design goals
Labels and Scales
None
None
Use labels to provide clarity
None
Pattern Matching
None
None
Correlation and Causation
None
None
None
http://tylervigen.com/spurious-correlations tyler vigen
The clustering illusion and confirmation bias
None
None
None
None
None
“That looks great!” - Our client
“That looks wrong” -Us
None
Apophenia
None
“I want this to be true and it looks like
it is” Confirmation bias
The clustering illusion
None
Case Study: Lunch attendance
THE PROBLEM:
HYPOTHESIS
AN HYPOTHESIS MUST…
AN HYPOTHESIS MUST… MAKE AN ASSERTION
AN HYPOTHESIS MUST… MAKE AN ASSERTION BE FALSIFIABLE
HYPOTHESIS HOW DO WE GET MORE PEOPLE TO SHOW UP
TO OUR MEETUP?
HYPOTHESIS PEOPLE ATTEND TOPICS THEY’RE INTERESTED IN
The whole story
Lets Remove the chart junk
TODO: BASIC ATTENDANCE
Maxima
Minima Maxima
TODO: ADD EXAMPLE OF ATTENDANCE
TODO: ADD EXAMPLE OF ATTENDANCE Frontend
TODO: ADD EXAMPLE OF ATTENDANCE Technology
CONSIDER THE VIEWPOINT
TODO: MEMBERS WITH ATTENDANCE
HYPOTHESIS PEOPLE ATTEND TOPICS THEY’RE INTERESTED IN
TODO: MEMBERS WITH ATTENDANCE Team Pizza
TODO: MEMBERS WITH ATTENDANCE Team Pizza Long tail
HYPOTHESIS PEOPLE ATTEND TALKS THAT THEIR FRIENDS ATTEND
TODO: MEMBERS GROUPED
TODO: MEMBERS GROUPED
TODO: MEMBERS GROUPED
TODO: MEMBERS GROUPED Friends
Conclusion
Consider your biases
Iterate
Let the data speak
Resources Edward Tufte Alberto Cairo informationisbeautiful.net
Thanks! Chris Keathley / @ChrisKeathley / c@keathley.io