Transcript SWEViZ - Nils Dickner

SWEViZ
IDEA AND GOAL
Visual representation of all traffic accidents in Sweden
between the years 2006-2010
Purpose was to do a exploratory analysis


Find interesting patterns and information
Create interesting hypothesis
DATA
Types of accidents

http://www.trafikverket.se/Privat/Trafiksakerhet/Olycksstatistik/

Separate data file for each county
Population (probability)

http://www.scb.se/
Problems

Adept the data to the visualization
REPRESENTATION AND LAYOUT
Choropleth map



Easy and effective representation
The user gains a better overview between
counties
InteractiveColorLegend work as a filter
Parallel Coordinates Plot



More detailed representation
The user gets a better insight
The headers work as a colormap filter
REPRESENTATION AND LAYOUT
Scatter plot 2D



The user can find relevant patterns
More detailed representation
Involve the user
Timeline



The user gets a easy overview for a specific county
Easy to understand and follow
The counties are prioritized
REPRESENTATION AND LAYOUT
Information tool



The user gets quick and relevant information
More specific facts about a county
Easy tool to use
Data tool



Involve the user
The user can get a better insight
Easy way to represent multivariate data
Conclusion
Possible hypothesis




More accidents with wild animals in counties with less
population.
Southern Sweden's worst drivers lives in Kronobergs län.
Drivers from north Sweden is the worst drivers in
Sweden.
It occurs more accidents during summer
Problems

Require more information to be certain, e g Statistics of each
road, weather, the environment etc
DEMONSTRATE APPLICATION
Q&A