Do Parking Ticket Quotas Exist? YES. An NYPD Traffic Ticket Time Series Analysis

Well, this past Tuesday was not my day. I woke up early in the morning to prepare for a morning phone interview, and while making coffee, I noticed that my motorcycle was curiously missing from where I had parked it the night before. I called 311 to see if it had been towed — nope. I called 911 to report it as stolen and file a police report, just barely making it back in time for my phone interview. With all the stress from a stolen motorcycle, I can’t say with confidence that the interview went all that great. I only found out after the interview that in fact it had been towed — the first (sort of) good news of the day.

Start of my tour to NYC from Detroit

I’m a new motorcyclist in New York — I’ve only had one in NYC for a little less than a month after driving it from Detroit. Due to a lack of a garage, or even a driveway (it is New York City, after all) I had been parking it on the sidewalk in front of my house, thinking I’d get a slap on the wrist before anything more drastic happened. Not the case with NYPD — I had to pay $200 to get it from their impound lot, another $115 for the illegal parking ticket, and finally $70 to replace the lock they broke to move it. I was told this was all a part of a crackdown by NYPD to systematically remove motorcycles parked on sidewalks. All this got me thinking —can I use NYC Open Data to predict where the best places and when the best times are to avoid getting a parking ticket? Are NYPD’s traffic ticketing patterns systematic, and can they be predicted? Is the rumor that traffic tickets are issued at the end of the month to meet certain policing “quotas” true, or merely myth?

My first step was accessing NYPD illegal parking data. Fortunately, Open Data NYC has exactly the data I need, and I was able to query illegal vehicle parking citations starting from this past week (June 2017) back to June 2010 (the ipython notebook I created to do all the data processing can be viewed here!) I subset this data further to just include citations for “Blocked Sidewalk”, and then plotted the sidewalk parking ticket counts across the entire city by day of week, date, and hour of day, as can be seen in the plots below. Middle of the week in the morning, especially on recent dates seems to be the most likely time for NYPD to issue parking tickets.

For the plots above, citations tend to increase overall with time since 2010, early in the morning (peaking at 9am), and drop off considerably on Saturdays and Sundays.

I also looked at the citations issued by community district. However, I wanted to control for the number of vehicles in each community district, as I didn’t want sparser community districts, such as those in the outer boroughs, to be underrepresented in my analysis compared to denser districts, such as those in Manhattan. I decided to look at the percentage of vehicles that were issued citations, out of all vehicles that belonged to residents of each community district. I used American Community Survey (ACS) data on population that has access to vehicles, which I was able to query from Enigma’s awesome Enigma Public data portal.

My ticket was issued at the green dot, right in the thick of the ticketing “red zone”

But are these daily, weekly, and yearly trends systematic? I next drilled down by community district to see if I could uncover any of these systematic time trends at a smaller geographic level. I looked at the top 10 community districts by % illegal sidewalk parking tickets issued, and used fourier analysis (used in applications ranging from helping astronomers identify pulsars to how shazam identifies songs) to see if I could uncover systematic temporal spikes in parking tickets issued across my 7 year period. Turns out, the answer is yes, there were systematic time patterns. The chart below shows the power spectrum chart, and interestingly, 8 of the top 10 community districts experience a spike in parking tickets issued once every 365 days.

I decided to drill down on Brooklyn Community District 2 and Brooklyn Community District 8, the two closest community districts to where my ticket and tow was issued, and also two of the highest in the city.

BK02 (left) and BK08 (right) time series aggregated to the 365 day level

It’s a little less obvious, but peaks occur around day 0 and day 360 for both districts. Day 0 in this case is June 9th, so both community districts have spikes once a year right at the end of May and beginning of June. Interestingly, there doesn’t seem to be any ticketing pattern with a monthly cadence, meaning it’s not as likely that the belief that a jump in tickets happen at the end of the month to meet quotas is true. But perhaps there’s a yearly quota that NYPD is trying to meet in a lot of these districts 🤔…

Trend analysis for community district s BK02 (left) and BK08 (right)— the highest and 10th highest %s in the city, respectively. Ratio is the % sidewalk parking ticket average for the last three weeks (per vehicle) divided by the first three weeks (>1 = increasing trend, <1 = decreasing trend)

I finally wanted to see if there was a systematic growth or decline in the issuing of tickets over the seven year period that the fourier transformation isn’t able to capture — the answer is yes, there is significant growth in the issuing of tickets, especially in BK02 and BK08. In fact, there’s been a 10 fold increase in the number of tickets in BK08 and over 20 fold increase in the number of tickets issued in BK02. The rumor that the NYPD has stepped up their aggression in issuing tickets appears to be markedly true.

To conclude, it looks like I would’ve been safest from a sidewalk parking ticket by parking in an outer borough (say, eastern Queens), in December, back in 2010. Of course, if I were living in an outer community district, I’d probably have much easier access to a driveway and a garage, completely circumventing my need to park on the sidewalk in the first place. Going forward, however, I’ll be sure to avoid the sidewalk in my neighborhood, especially in early June. NYPD does seem to ticket at a yearly cadence right around the end of May / beginning of June, and does seem to have really stepped up the ticket issuing since 2010.

The limitations of this analysis are many. I wasn’t able to get a truly accurate count of cars in each community district, as I have no way of accounting for changes in the count of car ownership across time (I only used 2015 ACS data), and I had no way of accounting for visiting cars (perhaps that’s why the lower east side in Manhattan had such high counts). I also had no way to account for driveway and garage ownership — I’d be interested coming up with some kind of google maps / computer vision algorithm, but I don’t quite have the time for that right now. One thing I could do is extend this to include all parking tickets, not just sidewalk parking tickets, so if I have more free time over the next few weeks, maybe I’ll throw something together. Stay tuned.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.