CollisionWatch is a visualization of traffic fatalities over one year across the United States. The data is, to say the least, sad, but putting these statistics into the context of when and where can lead to a better understanding of geographic, timely, and causal patterns. An increased understanding, and perhaps the revelation of predictable patterns and trouble spots can lead the way to better questions.
The data comes from the National Highway Traffic Safety Administration, used with permission, and denotes instances of automobile collisions in the year 2007 (US) where one of more fatality occurred, and what conditions were involved.
Because this data was rich with attributes, we were able to set up a filter set that could return some very specific results. For example, I can set the time of day and day of week to isolate workday commute times, or weekend nights, to see the picture change. Seasonal trends can be identified as well; cities encased in ice during the winter months have different patterns than you might expect –and each city is different.
The use of the “tri-state” filter enables the user to ask some very specific questions about factors that contributed to the incidents. For example: show me incidents where speeding was a factor but alcohol was not.
The basemap tiles are provided via the Bing Map Service and assembled within IDV’s VFX front end. The “Cartographic” map style is an interesting inverted treatment of the regular old road style, and invention of the clever folks at SoulSolutions, which we became aware of after we thought we had invented it. This darker, more muted basemap provides a much better foundation over which to drape data, showcasing the data as much more vibrant and identifiable.
This visualization uses Visual Fusion’s map viewer, timeline, and filter control. Incidents are shown as interactive objects on the map and in the timeline. Clicking any incident pulls up a basic report on the incident including contributing factors. The incidents are also represented as a hot-spot analysis map overlay. The hot-spot analysis, or heatmap, gives an immediate visual impression of clustered incidents and trouble spots where heavily stacked points may not. The combination of these visualization methods provides the benefits of insightful visualization and deep-dive reporting interaction.