The maps I made of NYC traffic-related fatalities in 2013 were recently featured in Brokelyn and BKMag. Given the interest in this data set and the discussions around VisionZero, I am going to dedicate a few more posts on the data set in the coming weeks.
Interestingly, the articles in Brokelyn and BKMag both focussed on Williamsburg. Yes, I did mention Williamsburg in my post. But as a statistician, the attention around that singular location made me a bit nervous. The number of fatalities compared to the size of the city is pretty small, so making conjectures about a single area can be a little tricky.
However, the size of the data set is much larger when looking at injuries as opposed to fatalities.
I have to admit, as a bike-commuter in NY, I was pretty stunned to see the map of cyclist injuries (reported to the police) in 2013:
Yes, each and every one of those dots is at least one injury. That’s over 3,800 reports in a single year! If you think about it for a while, this map is not surprising in a city with this many people. With over 8 million residents, 3,800 is a relatively small number. Even so, it does underscore how difficult VisionZero really is.
This map is so dense that it makes it difficult to detect any patterns. So I made a heat map of cyclist injuries here:
Williamsburg is still lit up there as a Hot-Spot. But, we are missing some important data here: ridership counts. Ask yourself this: if a route has 10 times the riders, but only 5 times the injuries as another route, should we consider it safer? That means that each individual cyclist that takes that route is only half as likely to get injured. Maps like the one above don’t account for ridership density.
That may explain why the East side of Lower Manhattan has so many more injuries than the West side:
The ridership coming off the bridge is huge, while there are no cyclists coming in from the West (as fun as it would be to cycle through the Holland Tunnel at rush hour)
Sadly, open data does not allow us to draw definitive conclusions about one neighborhood being more dangerous than another. We are just at the start of open data in New York City, and there are inherent limitations. Even so, these maps show that there are hot spots for injuries (not to mention fatalities) throughout the City. Putting ridership density aside, these maps reveal areas where the City should continue to invest in safety measures, in the spirit of Vision Zero.
Once again, thanks to The NYPD Crash Data Band-Aid for the data conversion. My data is only as good as the NYPD Crash Data Band-Aid’s data, so there could be some errors in it. This is a direct result of the NYPD choosing not to release the data in a more digestible form. Maps were done in QGIS, with Google Maps and HeatMap plug-in. Analysis done in IPython.Tweet
- emmadelosnardos likes this
- xbarbrastreisandx reblogged this from iquantny
- nmgreenfield likes this
- michael-xander likes this
- slumlordcitynyc likes this
- slumlordcitynyc reblogged this from iquantny
- jhrmn likes this
- alicescribbles reblogged this from iquantny
- nedsecondline reblogged this from iquantny and added:
- beigeinside likes this
- lawampm reblogged this from iquantny
- lawampm likes this
- aberjona reblogged this from iquantny
- wekeepeachotherhuman likes this
- iquantny posted this