Jason Holt bio photo

Jason Holt

Husband, father, stats wonk.

Twitter jason@arlingtoncrimemaps.com

Auto Crime in Arlington

Welcome to Arlington Crime Maps, a blog about crime and demographics in Arlington, Virginia. I hope to cover many topics, but for the inagural post lets explore auto crime in Arlington. I’d like to know when auto crime happens, where it happens, and any place or time with a greater than usual share of stolen vehicles.


Our dataset encompases 2,076 crimes in repoted in Arlington County from January 1, 2011 to December 31, 2013. Data were sourced from the Arlington County Police Department Daily Crime Report. Data attributes include a type of crime, a description, and block-level addresse (i.e. 1000 Wilson Blvd vs 1020 Wilson Blvd) where the incident occured, for all 2,076 crimes. Seven crimes have been removed from our dataset due to incomplete address information. I’ve hand annotated the crime data with neighborhood and police district identifiers, using the Arlington County 2013 GIS file.

For this post I’ll analyze the 364 stolen vehicles and 104 vehicle-related crimes (vandalism and larceny) that occured between 1/1/2011 and 12/31/2013. For this post I’m treating all vehicle related crime under the umberlla ‘auto crime.’

When auto crime occurs

We’ll start by looking at when and how often auto crime occurs. With this data we can say for sure if auto crime is going up or going down for a particular place or moment in time.

Auto Crime by Month

First we look at monthly auto crime totals for 2011-2013.

Auto crime by month

Looking at this chart a few things jump out at us immediately:

  • Auto crime for January 2011 is dramatically lower than the same period in 2012 and 2013. Possily due to our absurd snow totals at the time, or maybe a methodological error.
  • July, August, September, and November of 2013 all have much lower totals than the same periods in 2012 and 2011.
  • March, May, and October have nearly identical totals across all years.
  • Regardless of year, December is a strong time to have something bad happen to your car.

The most important thing to take away from this chart is that auto crime does vary by month, though there are some small patterns across years.

Crime by day of the week

Next lets look at variance between days of the week. I would naievely assume auto crime goes up on weekends, with more people parked at home.

Auto crime by month

It looks like my naive assumptions were half correct. Over our 3 year study period auto crime is generally static, with spikes on Friday and Monday. The Friday spike makes a little bit of sense, but the Monday spike is puzzling. So our day of week analysis is valueable by itself, even with some unanswered questions. The day of week may play a bigger role when combined with a geographic component, which we’ll need to keep in mind as we try and draw conclusions from the maps.

Where auto crime occures

Now that we’ve got a better understanding of when auto crime occurs, let’s look at where it occurs. Our primary method of visualization will be a Heat map. A heatmap represents intensity with color, in our case the number of auto crimes in a particular location. We’ll overlay the heat map on an actual map of Arlington to complete the representation.

One note on the geographic representation: crime data is provided in a block-level format (2200 Clarendon Blvd vs 2205 Clarendon Blvd), but is sufficently accurate for our purposes.

Heatmaps of auto crime

Our first heatmap is of all auto crime in the data set:

Auto Crime in Arlington VA 2011-2013

We see two interesting concentrations of auto crime, one between Clarendon-Courthouse and Ballston, and the other around the Pentagon City Mall. Both are dense population-wise, and the Pentagon City is a huge retail center that attracts lots of cars. Smaller concentrations near the south-eastern tip of Bailey’s Crossroads and 395 and S Glebe are also interesting.

Yearly Heatmaps

Now lets look at the same map, but showing auto crimes by year.




Breaking out the heatmaps by year gives us some insight that the full heatmap couldn’t do alone. The Pentagon City Mall cluster is strong each year, while the Clarendon-Courthouse blob shifts towards Ballston after 2011. Auto crime in the Bailey’s Crossroads cluster grew substantially in from 2011 to 2012, and became a cluster in its own right in 2013. All the major 2013 clusters are shopping or population centers, which you would expect to attract larger than normal numbers of cars, which probably attracts more auto crimes.

Pentagon City-Crystal City Detail Map

One of the strongest clusters in the heatmap across all years encompasses Pentagon City and Crystal City. Lets use a heatmap of auto crime in only those areas to see if incidents are widely disprsed or clustered in any particular area.

Auto Crime in Pentagon City-Crystal City 2011-2013

Our answer is clear: auto crime is clearly clustered in the Pentagon City Mall parking areas. A ParkoPedia Map for the mall estimates the number of parking spaces at 5168 spots. Anyone who makes the mistake of, say “running to Striderite really quickly for shoes for the kid,” on any given weekend, can tell you securing a spot can be a dangerous and dramatic affair. It is no surprise that many cars means more auto crime for the area

What is a surprise is that 21% (99 count) of all auto crime in Arlington County in 2011-2013 occured in or around the Pentagon City Mall area.


So what did all these charts tell us about auto crime in Arlington? We’ve learned:

  • Auto crimes tend to occur in popular neighborhoods, like Clarendon and Ballston.
  • Auto crime in Bailey’s Crossroads and Ballston has gotten somewhat worse in 2012 and 2013 than 2011.
  • One in every five incidents of auto crime happens in or around the Pentagon City Mall parking lot.

Thanks for reading, and please feel free to leave questions or comments below, or via email to jason@arlingtoncrimemaps.com.