Data Visualization – What is data visualization? – Why are visualizations useful?

View Github Repository Open presentation in a new window

mkfreeman

See all presentation from mkfreeman

Data Visualization – What is data visualization? – Why are visualizations useful?

0 0

db_lecture

On Github mkfreeman / db_lecture

Data Visualization

Michael Freeman

Institute for Health Metrics and Evaluation

Outline

  • What is data visualization?
  • Why are visualizations useful?
  • Case study exercise
  • Databases

What is data visualization?

Are these data visualizations?

Immersion Media

What is data visualization?

The practice of representing quantitative information through graphical encodings

Why are visualizations useful?

Understanding data

  • Mean x: 9.00
  • Mean y: 7.50
  • Variance x: 11.00
  • Variance y: 4.12
  • Correlation between x and y: .816
  • Linear regression: y = 3.00 + .500x

Communication

Florence Nightengale

“to affect thro’ the eyes what we fail to convey to the public through their word-proof ears.”

Leveraging pre-attentive cognition

How many 3's are there?

How many 3's are there?

Case study: U.S. health measures by county

Data

Prevalence of health risks (smoking, obesity, etc.), vary by:

  • Location (county, state, national)
  • Time (1995 - 2012)
  • Sex (male, female, both)

Who will use this information?

  • County/state level officials
  • Insurance companies
  • Non-profit organizations
  • Researchers
  • General public
  • Tobacco companies (?)
  • You

What questions do they have?

  • What are the trends over time?
  • Are there geographical health patterns? What are they?
  • Which locations have improved the most over time?
  • How does my county compare to my state, and the nation?
  • How do men and women compare?
  • Are there relationships between health burdens?

How can these questions be answered graphically?

  • Time trends: lines
  • Geographical patterns: map
  • Best/worst performers: spatial position
  • State/national comparisons: reference lines
  • Link

Databases

Why visualizations need databases

  • Lots of data
  • Only a fraction is shown at once
  • Need a mechanism for selecting and filtering the data

Database querying format

SELECT var1,var2,varN FROM database.table WHERE var1=value1 AND var2=value2

Controlling queries in visualizations

SELECT topic,sex,year,value FROM health_db.us_data WHERE topic="smoking" AND sex="female" AND year=2012

Summary

  • Data visualizations as diverse graphical representations of information
  • Importance of leveraging pre-attentive cognition for understanding and communication
  • Endless applications
  • Need databases to make (scalable) applications

Contact

Michael Freeman

mikefree@uw.edu

@mf_viz

http://mkfreeman.github.io/

GH590H: Interactive Data Visualization