How to make inspiring data visualisations (Jan Willem Tulp)

• http://tulpinteractive.com

General

• Apartment or Car market websites:
  • It's easy to find something exact, but
  • If you don't know what you want, the UI doesn't help you.
• How do we get the overview of a data set?
• Date -> Concept -> Visual design + Presentation
• Use some of the properties of the data to help the visualisation:
  • E.g. Iraq death toll bloody barchart

• What do people find interesting?
  • Different between people
  • Changes with time

• Visualisation magic quadrant
  • Common vs new/novel
  • Comprehensible vs incomprehensible

• Martini glass model
  • Start simple and focused, then open up

• Trigger -> Action -> Variable reward -> Effort -> Trigger
  • Something initially attractive,
  • Ability to interact and discover by yourself, but with a bit of work

• Tools:
  • Data preparation with Python
  • For the visuals: frameworks like D3

Dutch Archive of Architecture and Urban Planning

• What does the archive look like?
• Visualisation design = Finding a visual representation of a data set that works for a particular situation.
• Visualising the data early helps to understand how does the data look like and what is its structure. It also shows what works and what looks nice.

Hipparcos and the star catalog

• https://en.wikipedia.org/wiki/Hipparcos
• http://sci.esa.int/star_mapper/

Tree density visualisation for Nature

• Finding the right metaphor so that visualisation looks like forests.

Dutch voting of 2012

• Which cities vote in a similar way?
• What can we see from this visualisation?

Tools

• Frog - an alternative to NLTK, specifically for Dutch.