NYC emergency response time correlates with small crime incidents

To follow up on this article or ask any questions, you may email the team directly. All e-mail addresses are @stern.nyu.edu. Direct your email to Dylan Riley (dylan.riley@), Chelsea Rosetti (cer312@), Thomas Pauwels (ttp217@), or Ashley Kaufman (ahk291@). We apologize for the inconvenience of typing out email addresses, but there have been some incidences of virus/spamming email with mailto links under this directory. We welcome your questions!

Graffiti occurrences per thousand, by borough, compared to average emergency response times: little or no correlation

The study began by considering data sets involving the prevalence of graffiti and medical and fire response time. The team wanted to uncover whether a quicker response time from authorities meant less small crime. Using a per thousand people statistic to see how the number of graffiti occurrences varied by borough in comparison to medical and fire response time, the team found a weak relationship between low response time and low per capita graffiti occurrences.

In the graph below, the graffiti occurrences were divided by borough population (and multiplied by 1000, to get a useful number).

So is a speedy emergency response time all that is needed to ward off low level crime? Not likely. The relationship is very weak, though more plausible when fire response time is considered. While the Bronx has the highest per capita statistic, it is average in both fire and medical response times. All was not for naught, however; this disproves the generalization that a speedy emergency response time means less small crime. The group wasn't satisified, though, a new metric was considered.

Risk-averse street artists? Statistics suggests it.

The team was also interested in seeing how graffiti occurrences varied in boroughs, given the standard deviation of the response times of the emergency authorities in the boroughs. The results of this experiment were fairly surprising, showing that graffiti artists may be risk-averse.

Here is how the team arrived at this conclusion. First, the occurrences of graffiti in each borough were counted. Then, the standard deviations of the fire and medical response times in each borough were calculated. Graffiti artists tended to work more in boroughs that had less variance in response times from emergency authorities. Do graffiti artists operate more when they have a better feel for how quickly authorities will respond? Perhaps so. Perhaps not. This relationship is statistically significant, however, since both datasets consist of well over 10,000 samples.


The obvious exception here is Manhattan, which has the second least number of graffiti incidents while it has the second least standard deviation of authority response. This could be due to the fact that Manhattan is the most densely populated of the boroughs and more easily policed. It could also likely be a consequence of numerous other demographic and architectural factors that make Manhattan less attractive for graffiti artists, relative to the other boroughs.

Update: Seasonal cycle shows a stronger correlation between emergency response times and graffiti occurrences

It seemed odd that small crime had no correlation with emergency response time, so the team investigated a little further and found some new results. The group's new query counted graffiti occurrences and arranged them by borough and month of occurrence. This result was compared against a similar query analyzing total average emergency response time in borough, according to month.

Graffiti occurrences throughout the year

Emergency response times throughout the year

This graph includes the first and last month of each year, along with the month with the longest response times: July. Indeed, July has the most occurrences of graffiti as well throughout New York's boroughs. This might be attributed to kids being out of school for the month of July.


To verify these findings, feel free to download the Access file here.


Download the excel document with summarized database results here. This data has been queried and filtered from massive databases at the NYC mine.

You can access the NYC raw data by visiting the NYC mine website. Our data was extracted from the following sets:Prevalence of graffiti and medical/fire response time.

Images for background from Google Images:

Graffiti: Mark Ross, Canned Goods

Ambulance and fire trucks: City of Kalispell, Montana, Fire Dept, Edmonton-all Flickr album. Images will be removed immediately at the request of the rights holder. Contact: dylan.riley (at) stern (dot) nyu (dot) edu.