arlingtonArlington County has released its own set of bicycle accident data. There were 227 incidents recorded from January 1, 2010 to December 31, 2012 (though the first recorded accident wasn’t until March 1, 2010). That’s 6.3 reported incidents a month. I’ve made a map for them, the Arlington Bicycle Accidents Stat Mapper. It’s the same program as the DC Bicycle Accidents Stat Mapper.

I did not combine the two jurisdictions because the date ranges are different, and because the data formats were different.

Arlington’s map has 174 locations. The spot with the most accidents was Lee Highway at Fort Myer Dr, with 8 accidents recorded there. Clicking on a pin will show detailed information for all of the incidents at that location.

The source data has an “at” street and a “cross” street for each record. Some of them have a “landmark” field which is often used to describe the block number, and another field (3, actually, which I’ve combined) is sometimes used to describe the distance from the intersection.

The geocoding process was a bit problematic, as always. 69 of the records already had latitudes & longitudes. To get coordinates for the others, I geocoded the two cross streets, then manually adjusted obvious errors.

For this version at least, each pin shows all records that share the same cross streets in the same order. Thus, S Four Mile Run Dr & S Walter Reed Dr is a separate pin than S Walter Reed Dr & S Four Mile Run Dr. When multiple pins have the same coordinates, you can see only the one that happens to be drawn on top. For this particular intersection, I slightly moved one set of records so both pins are visible when you zoom in. There might be other intersections where two versions of an intersection are overlaid. Viewing a heat map instead of pins help to address this problem. (There are better solutions which I will try later.)

There was another troublesome intersection, which was listed simply as 31ST ST N intersected with “DR”. The geocoder mysteriously placed that in DC, so I manually moved it to a random section of 31st St in Arlington. I thought it was better to have it in an inaccurate position than to remove it from the dataset.

As with DC’s bike accidents map, you can’t really use the map to judge the most dangerous locations, since we don’t also show traffic levels. You can see most accidents happen along Wilson Blvd and in Rosslyn. There’s also a high-incident area around Four Mile Run and Walter Reed Dr, which is where the only bicycle fatality occurred.

A few summaries: the most common weather recorded was “No Adverse Condition,” in 94% of the cases. Rain was recorded for 5% of the accidents. The road condition was dry 93% of the time, wet 7% of the time. The data has a direction field and an “action” field for two vehicles. Vehicle 1 and Action 1 refer to the at-fault party. Not sure if there’s a way to confirm that “party 1” refers to a car. The action field for vehicle #1 is “Did Not Have Right-Of-Way” 34% of the time. The action field for vehicle #2 is “No Improper Action” 83% of the time.

The data is interesting to review, and it’s good to have it made public. Hopefully this will help make everyone more mindful of the need to safely share our public streets.

Arlington Bike Crash Data

3 Responses to “Arlington Bike Crash Data”