Data Visualization - University of Iowa · Data Visualization Principles Bar charts Often the best choice Good for comparing categories Good for showing data over time. 5/6/2019 7
Post on 27-May-2020
5 Views
Preview:
Transcript
5/6/2019
1
Data Visualization for Infection Prevention
Jorge Salinas & Jeff KritzmanUniversity of Iowa Hospitals & Clinics
Agenda
Introduction & Motivation
Visual Perception & Memory
Data Visualization Principles
Data Visualization for Infection Prevention
Introduction & Motivation
5/6/2019
2
Aesthetics matter
Aesthetics matter
Case study
5/6/2019
3
Adding a trend line
A different trend line?
Design choices matter
5/6/2019
4
Standing caveat
“It depends”
https://marketingland.com/stop-saying-depends-give-better-answers-147113
“The visual display of quantitative data to facilitate understanding”
Stephen Few & Andy Kirk
What is data visualization
“The visual display“The visual display of quantitative data“The visual display of quantitative data tofacilitate understanding”
Stephen Few & Andy Kirk
“The visual display of quantitative data tofacilitate understanding”
Stephen Few & Andy Kirk
Visual Perception & Memory
5/6/2019
5
Visually encoding for rapid perception
People naturally look for patterns
Easier to compare numbers visually
Design with this in mind
How many 3’s?
FBICIANSAFBI CIA NSA
Short term memory test
What does this say…
5/6/2019
6
Memory
Short term memory is limited
5-9 ‘chunks’
FBICIANSA vs FBI CIA NSA
This is why data visualization works
Compare numbers without relying on memory
Data Visualization Principles
Bar charts
Often the best choice
Good for comparing categories
Good for showing data over time
5/6/2019
7
Flipping and sorting bar charts
Flip itEasier to readDon’t need to turn head
Sort itCompare relative performance more easily
Line charts
Good choice for showing data over time
Numbers:Use bar chart
Rates:Use line chart
Scatter plots
Great way to compare two variables
5/6/2019
8
Heatmaps
A good way to see patterns
Compare categories
Compare time periods
Small multiples
Pie charts are hard to read
5/6/2019
9
3-D charts are hard to read
A
B
C
0
1
2
3
4
5
6
7
8
a b c d e
A B C
Keep it simple
Reduce clutter
Maximize value-added pixel ratio
Watch out for “chart junk”
Grid lines
Borders
Background fill
Images
Excessive labels A
B
C
0
2
4
6
8
a b c d e
43
6
32
3
5
3 32
7
5
3
87
A B
Grid lines
Avoid or make very light
Use target or goal lines instead
5/6/2019
10
Labels
Can be helpful
Often obscure meaning
Label alternatives
Label the last point
Include a small table
Color
Use only when it adds meaning
5/6/2019
11
A little color goes
a long way
Picking the right metrics
Outcome Metrics
What you are interested in
Process Metrics
Measure the process
Influence outcome measures
Include both in reports
Data Visualization for Infection Prevention
5/6/2019
12
Surgical site infections (SSIs)
SSIs by team
Team
1Te
am 2
Team
3
SSIs by surgeon
5/6/2019
13
Intra-oppatient temperature
Surgical process metrics
Integrated SSI dashboard
5/6/2019
14
Hospital-acquired infections (HAIs)
HAIs by location
Further reading
Edward Tufte
“The Visual Display of Quantitative Information”
Stephen Few
“Information Dashboard Design”
5/6/2019
15
Thank you!
Jorge Salinas jorge-salinas@uiowa.edu
Jeff Kritzman jeffrey-kritzman@uiowa.edu
University of Iowa Hospitals & Clinics
@jlsalinas7 @jeffkritzman
www.InfectiousDataViz.com
All images from Pixabay, Unsplash, StockSnap.io, or taken/created by us, unless otherwise noted
top related