Big lists of crashes are difficult to turn into something to make decision upon. When we map crashes, and present relationships between near crashes, such as listing other nearby crashes, we can start to understand a bit more about the context of a single crash. Being able to view the crash site enables exploration of why an accident might have occurred. Exploring this data in context could allow for more targeted capital investment, ultimately making the roads safer for us all.
Data was sourced from VicRoads, massaged in PostGIS and when it was determined that it would be too big to run client-side, it was loaded into CartoDB. CartoDB's version of Leaflet is used, as well as Google's API. (Yep, that's right, this map use two mapping APIs!)
Zoom into the map to start seeing individual accidents.
When you're zoomed out, you can view a density map, showing relative numbers of crashes in an area.
When you're zoomed in, you can start to see actual accident locations, and these are coloured by severity.
Clicking on a crash delivers more information, such as what type of vehicle was involved and the date.
Also, when you have clicked on one crash, the map will list 10 crashes that are within 250 m of the crash, ordered by distance.
Crashes can be explored in terms of detail, but also in context with the Google StreetView image.
Crashes can be linked through from StreetView or from the table, not just from the map.