Luopan 罗盘

New York City is a complex place to drive.  And when it comes to parking, there are plenty of rules and regulations to follow.  It’s no wonder that sometimes people get confused and end up getting their cars ticketed or towed. 

But in all of these rules, there is one thing that very few drivers seem to know. As of late 2008, in NYC you can park in front of a sidewalk pedestrian ramp, as long as it’s not connected to a crosswalk.  It’s all written up in the NYC Traffic Rules, and for more detail, take a look at this article. The local legislation making these parking spots legal was proposed by Council Member Gentile, and adopted by the Department of Transportation before it ever made it for a vote.  Though few people seem to know about the change.

Is it a problem that drivers don’t realize that there are some extra parking spots they are now allowed to park in?  Not so much.  But, I’ve got a pedestrian ramp leading to nowhere particular in the middle of my block in Brooklyn, and on occasion I have parked there.  Despite the fact that it is legal, I’ve been ticketed for parking there.  Though I get the tickets dismissed, it’s a waste of everybody’s time. And that got me wondering- How common is it for the police to give tickets to cars legally parked in front of pedestrian ramps?  It couldn’t be just me…

In the past, there was not much you could do to stop something like this. Complaining to your local precinct would at best only solve the problem locally.  But thanks to NYC’s Open Data portal, I was able to look at the most common parking spots in the City where cars were ticketed for blocking pedestrian ramps.   It’s worth taking a moment upfront here to praise the NYPD for offering this dataset to begin with.  Though we are behind on police crime data in the city, we are ahead in other ways and the parking ticket dataset is definitely one of them.  

What I found when I dove into the data surprised me.  To start, I found the top address where this ticket were given: in front of 575 Ocean Avenue in Brooklyn, where over $48,000 in parking fines were issued in the last 2.5 years. 

To my surprise, the spot, (or really spots since there are two ramps), are legal, since they are in the middle of the block, with no crosswalk.  $48,000 in tickets at a legally parked spot, and that is just the last 2.5 years.  If we had more historical data, I suspect we would see a similar story. 

The next top spots on the list had the same story to tell.

1705 Canton Avenue in Brooklyn, 273 Tickets, $45,045: Legal.

270-05 76 Avenue in Queens, 256 Tickets ($42,440) Legal.

143-49 Cherry Ave, Queens, 246 Tickets, ($40,590).  Legal.

Those were of course the most extreme cases that topped the list.  I started to skip down the list.  This spot in Battery Park, ranked #16 on my list and the top spot in Manhattan, had 116 tickets ($19,140) and turned out to be legal.  (Yellow paint has no meaning in NYC traffic enforcement, and the spot has no crosswalk)

I started to skip down the list faster and faster.  Take #1000 in my top list, at 1059 Virginia Avenue, where 8 tickets had been given ($1320).  It is a classic T intersection, meaning it’s legal.

See the pattern?  I did.  So I then selected 30 random spots that had received 5 or more tickets over the time period, and based on Google Maps found that all of them appeared to be legal parking spots!  (Randomly selecting spots with a single ticket in the database showed some illegal spots as well, so I chose 5 as a conservative cutoff.)

How many spots received 5 or more of these pedestrian ramp tickets in the last 2.5 years?  We are talking 1,966 spots that are generating about 1.7 million dollars a year in tickets at parking spots that are mostly legal.  Are all 1966 spots legal?  Surely not, but the majority sure are and many more that have fewer than five tickets are likely legal too.  

To see where the hotspots were for this sort of abuse of drivers, I mapped the top 1000 pedestrian ramps generating parking tickets below, so you can see if you have a spot in your neighborhood that the police have been wrongly ticketing.  Clicking on a spot tells you if the tickets are given in front of, or opposite the address, as well as how many tickets have been given.  

My challenge to you…  click on a spot and then look at pictures of it on Google Maps to determine if its legal.  You will most likely find it is. 

Other findings in the data:

Brooklyn’s 70th Precinct seems to have the most cars wrongly ticketed, bringing in over $100,000 in fines a year.  The 77th, also in Brooklyn, comes in second:

So what is going on here? Well, it seemed from my browsing that many police officers were systematically ignoring the 2009 rule change, all across the city.  

But that is not where the story ends. As mentioned, we live in a city progressive enough to release parking ticket data.  I’ve repeatedly said that putting data into the hand of citizens will make our city run better and more equitably.  So, I reached out to the NYPD via the Mayor’s Office of Data Analytics to tell them about the findings.  They helped me get in touch with the appropriate people at the NYPD, and then I waited.

After a couple of weeks, I received the following quote from the NYPD:

“Mr. Wellington’s analysis identified errors the department made in issuing parking summonses. It appears to be a misunderstanding by officers on patrol of a recent, abstruse change in the parking rules.  We appreciate Mr. Wellington bringing this anomaly to our attention.

The department’s internal analysis found that patrol officers who are unfamiliar with the change have observed vehicles parked in front of pedestrian ramps and issued a summons in error. When the rule changed in 2009 to allow for certain pedestrian ramps to be blocked by parked vehicles, the department focused training on traffic agents, who write the majority of summonses.

Yet, the majority of summonses written for this code violation were written by police officers. As a result, the department sent a training message to all officers clarifying the rule change and has communicated to commanders of precincts with the highest number of summonses, informing them of the issues within their command.

Thanks to this analysis and the availability of this open data, the department is also taking steps to digitally monitor these types of summonses to ensure that they are being issued correctly.”

I was speechless.  THIS is what the future of government could look like one day. THIS is what Open Data is all about.  THIS was coming from the NYPD, who is not generally celebrated for its transparency, and yet it’s the most open and honest response I have received from any New York City agency to date. Imagine a city where all agencies embrace this sort of analysis instead of deflect and hide from it.   

Democracies provide pathways for government to learn from their citizens. Open data makes those pathways so much more powerful.  In this case, the NYPD acknowledged the mistake, is retraining its officers and is putting in monitoring to limit this type of erroneous ticketing from happening in the future.  In doing so, they have shown that they are ready and willing to work with the people of the city.  And what better gift can we get from Open Data than that. 

—-

When’s the last time someone asked you where you lived or worked, and you answered by naming a police precinct?  I’m guessing never.  Why?  Because while police play a vital role in our communities, we tend to connect more to our neighborhood boundaries than our precinct numbers.  (How many of you can even name the number of your home precinct?)

And yet, for as long as modern crime statistics have been recorded here in our city, the police have been delivering those statistics to us, aggregated as counts by precinct number.  This type of aggregated data is useful to the NYPD, as it allows them to move resources around according to crime and assess progress, precinct by precinct.  But unfortunately, it’s not so useful for the rest of us. Single precincts can cover a wide range of neighborhoods, so even when you hear there is a change in crime levels in your precinct, you might not know if it affects a specific neighborhood.  You might want to split crime data up in other ways, like by street or council district, and that has not been possible.  Or you might want to know on what days or times of day crimes in a neighborhood occur most often, but with aggregate date, no luck.

But something big quietly happened last week; something that starts to put some of the issues I just outlined to bed.  For the first time in its history, the New York Police Department released incident-level reported crime data. That means for each individual reported felony (for the first three quarters of the year), the citizens of New York now have access to data showing where and when that felony was reported instead of a single number telling us the total number that were reported in a precinct over a period of time. There are countless insights to derive within, and with this post I am going to start to scratch the surface.

(Note that location data was only released for six out of seven of the major felony categories-  rape locations have been left off to protect the privacy of victims. So for the rest of this post, I’ll talk about the Six Major Felonies, which are the “Seven Major Felonies” that NYC reports, minus rape.) 

Before I dive in, it’s worth congratulating the NYPD and the de Blasio administration for making this happen.  It’s been a long time coming and I have no doubt that some innovative techniques and insights will arise from this release.  Those new ideas will eventually find their way back to our government and make our city even safer. But it’s also worth noting that many other cities (e.g. Los Angeles, Chicago, Seattle, Denver, Boston) provide more granular crime data, with longer histories (many years, instead of just three quarters of data), a broader set of reported crimes (not just six or seven major felonies), and with more detail (has an arrest been made, what are the victim demographics), so as exciting as this moment is, it’s good to remember that New York is still very far behind its peers on police data transparency.  

OK, enough context. On to what I learned:

Mapping and Ranking NYC’s Reported Felonies by Neighborhood

The data released is from the first three quarters of 2015, and gave location and time for six major felony categories: Robbery, Murder, Grand Larceny Auto, Grand Larceny of Motor Vehicle, Burglary and Assault.  We have had some view on this data in the past – a few years back the NYPD put out its Crime Map, but did not provide access to the underlying data, a BIG Open Data faux-pas.  One of the provided map views shows aggregate crime data by precinct, but I chose to redo this sort of analysis at the neighborhood level (well at least Neighborhood Tabulation Area level NYC’s census designated neighborhoods), which is much more granular than precincts.

For each neighborhood, I ranked it against the 188 neighborhoods with NTA’s (removing parks, cemeteries, airports and Riker’s Island).  I created a rank for each of the six felony types, where a rank of 1 means that the neighborhood leads for the most of that felony type per resident, and a rank of 188 means the neighborhood has the least of that reported crime per resident.  Lastly, I calculated which neighborhood had the highest average ranks versus the lowest.  This, in essence, gives each neighborhood an overall felony crime rank, giving equal weight to rankings from each type of crime.  (You could argue that weights should be calibrated differently, but we’ll leave that for a future post.)  With that, I labeled each neighborhood with its overall felony rank and made the map below:

Clicking on any neighborhood will show you its rank for each of the six felony types, as well as its overall rank (which is also what it is colored by).  

Interestingly, the safest seven neighborhoods by this measure are all contiguous in Staten Island and are all fairly low density.  It’s also interesting to see that Midtown-Midtown South is ranked 1st overall.  This is likely due to the low number of residents in that area compared to the number of people, since each crime rate here is normalized by the population size of the neighborhood.

The top 10 ranking neighborhoods for all felony types can be found in the table below:

While looking at the data, I also noted a few more interesting patterns…

More Robberies Happen in the three hours between 2-5PM, than 2-5AM

One great advantage of disaggregated data is that we can start to explore the time of day that different crimes happen.  I know many a New Yorker who is nervous about getting mugged coming home from a late night out.   The city certainly feels less safe late at night.  But I was interested to learn that the newly released data shows that robberies happen throughout the day, not mainly in the late evenings.

It turns out there are more robberies between 2-5PM then 2-5AM. This fact seems largely due to a large spike on weekdays that begins right after 2PM, seen in the table below, possibly coinciding with school release times, but obviously this warrants much deeper analysis.  (Continued investments in after school programs anyone?)  Of course there are more people out during the afternoon than early morning, so that does not mean any individual is more likely to be robbed when they are out and about.  But nonetheless, it’s an interesting finding which challenged an assumption I had made about New York City.  That being said, the most problematic single hour of the week (as opposed to hour of day) seems to be between 4AM and 5AM on Sunday mornings (e.g. late Saturday night), not surprisingly right after all NYC bars close.

Murder and Assault are at Least Four Times More Common in Our Least Affluent Neighborhoods as Compared to Our Most Affluent Ones.

I put each neighborhood into one of five buckets, 1 being the lowest 20% median income income quantiles (those neighborhoods with a median income of less than $34,000) and 5 being the highest (those neighborhoods with a median income of $68,950).  I then calculated the felony rates for the residents of each of the five buckets.    

For example, the first number in the burglary row is 1.16 meaning that in the least affluent 20% of neighborhoods, the number of burglaries per resident is 1.16 higher than the city average.  

Five out of six types of major felonies are more common in less affluent neighborhoods than more affluent ones, as seen by the general lowering of numbers from left to right in the table.  The exception is Grand Larceny, which is 1.75 times more likely in the most affluent 20% of neighborhoods.  The 1 to 5 Ratio numbers on the right of the table tell you how much higher the reported felony rate is in the least affluent 20% versus the most. The numbers that stand out are Murder and Felony Assault, which are more than 4 times more likely in less affluent neighborhoods.  Of course it’s no surprise that crime is higher in less affluent neighborhoods, but quantifying it this way shows just how much in a more digestible form.  This was not generally possible to understand before last week’s open data release, as census data is not usually collected by police precinct.

Conclusion

There is so much more to do with this data, but in order to do a meaningful analysis, the public needs access to more than one year of crime data, and other reported crimes besides major felonies. As noted, many other cities are providing just that.  Just today, the de Blasio administration proudly proclaimed that 2015 was “the safest year in the modern history in New York City”, and yet they are holding back the very data that can back up that claim with more specificity.  That is truly unfortunate, but the good news is we are seeing transparency for the first time in our history .  And that is a win for all New Yorkers, including those that work at the NYPD.  Thanks again to all that made this happen (with a special thanks to The Transparency Working Group,  The Mayor’s Office of Data Analytics and DoITT who worked behind the scenes).

—–

Match #170

Anne-Marie Van Dijk by Viviane Sassen for Wallpaper* Magazine May 2009 | The Darvaza Gates Of Hell crater in the Karakum Desert, Turkmenistan photographed by Amos Chapple

It’s well known that in Manhattan, the Avenues go north-south and the Streets go east-west. But what about the Roads, Places, and Drives? And what about in other boroughs?

Unlike like Emperor Kuzco, I was actually born with an innate sense of direction.  If you’re like me, and you use the Sun to navigate, you probably appreciate cities with gridded street plans that are oriented in the cardinal directions. If you know that your destination is due west, even if you hit a dead end or two, you’ll be able to get there. However, not all urban planners settled on such a simple layout for road networks. For some developers, topography or water may have gotten in the way. Others may not have appreciated the efficiency of the grid. This visualization assesses those road networks by comparing the relative degree to which they are gridded.

To generate the graphic, I first calculated the azimuth of every road in ten counties (plus one parish and D.C.). I tried to choose consolidated city-counties to keep the focus on urban centers, but for larger counties, I opted not to clip the shapefile to the city boundary. All calculations were made in a sinusoidal map projection using the central longitude of the area of interest. I then graphed the angles on rose diagrams (wind roses) using bins of 5° to show relative distributions for each area. The plots were scaled such that the maximum bar height was the same on each rose. To ensure rotational symmetry in the plots, each azimuth was counted twice: once using the original value and once using the opposite direction (e.g., 35° and 215°). As such, all streets, regardless of one-way or two-way traffic, were considered to be pointing in both directions.

The plots reveal some stark trends. Most of the counties considered do conform to a grid pattern. This is particularly pronounced with Chicago, even though much of Cook County is suburban. Denver, Jacksonville, Houston, and Washington, D.C., also have dominant grid patterns that are oriented in the cardinal directions. While Philadelphia and New York are primarily gridded, their orientations are slightly skewed from the traditional N-E-S-W bearings. Manhattan is particularly interesting because it has a notable imbalance between the number of streets running the width of the land (WNW to ESE) and the length of the land (NNE to SSW). New Orleans and San Francisco express some grid-like forms, but have a nontrivial proportion of roads that are rotated in other directions. Downtown Boston has some gridded streets, but the suburban grids are differently aligned, dampening the expression of a single grid on the rose diagram. Finally, the minimal geographic extents of the grids in Charlotte and Honolulu are completely overwhelmed by the winding roads of the suburbs, resulting in plots that show only slight favoritism for certain street orientations.

If you want to see more detail, a full-resolution version of this graphic can be downloaded here:

https://www.dropbox.com/s/my7y24hrzvhagce/Road_Orientation.png

Data source: http://www.census.gov/cgi-bin/geo/shapefiles2013/main