Can interactive data visualisation help us communicate uncertainty?

Robert Grant, Kingston & St. George's

Interactive

Why?

  1. To give an overview
  2. To convey scale and complexity
  3. To explore the data and comparisons therein
  4. To communicate information
  5. To tell a story
  6. To attract attention and stimulate interest

(objectives proposed by Gelman & Unwin 2013)

Conclusion: one image is not enough

Example one

goo.gl/ykBLke

Interactive

Example two

stateofobesity.org/adult-obesity

Interactive

Anatomy of webpages

JavaScript libraries

A D3 page

Or avoid the JS entirely

Low interactivity

goo.gl/xltEzd

Moderate interactivity

square.github.io/crossfilter

Extreme interactivity?

sumsar.net/best_online

Extreme interactivity?

lock5stat.com/statkey

Interactive

Example three

goo.gl/zKSFrs

The challenge

Example four, part 1

Example four, part 2

wap.weather.com/news/weather-hurricanes/sandy-path-evolution-20121026

Some uncertainty options

Experiment one

Considerations


robert.grant@sgul.kingston.ac.uk
www.robertgrantstats.co.uk

/