How to Make NYC a Safer City for Drivers
Why does analyzing NYC collision data have value? What is at stake?
Our objective in working with the New York City collision data was to examine trends in the data to yield actionable insights with respect to how to reduce collisions in NYC. NYC is a very unique city, characterized by extremely high population density and a high proportion of public transportation/rideshare vehicles/taxis as compared to private cars. As a result, the city’s data is not easily extrapolated to other cities and it is necessary to delve into NYC’s own data to figure out where improvements can be made as it relates to driver safety. Vehicle accidents are one of the leading causes of death in the world, making the analysis of driver safety a crucial objective for NYC to strive for.
Thesis: By using open source traffic data as is provided by the NYC government, there are actionable insights to be yielded that can be used to improve upon traffic safety in New York. Traffic safety cannot be chalked up completely to chance; the data revolution of recent years has afforded us the opportunity to gain a strong understanding of why events occur, and the extensive datakeeping of the city allows us to explore. This analysis explores how data methods on an extensive database of recordkeeping can provide a deeper understanding through which governments can form policy.
Even while some of the conclusions related to collisions might be seen as obvious, it is important to verify intuition when forming policy; adequate policy objectives cannot be formed on hunches and assumed truths. The recommendations below dive into areas of focus based on confirmed conclusions from data of the past five years.
Actionable Insights / Recommendations
Distraction is the #1 Reason for Traffic Collisions
“Driver inattention/distraction” is easily the leading cause of accidents (if you filter out the unspecified causes). While still a rather broad collision cause, there are notable areas of distraction that can be pointed out. Phones, for example, are a major distraction for all people and at least somewhat responsible for many accidents. Whether it be calling or texting, cell phones are becoming increasingly distracting, especially when considering the extensive communication needed for work in a city as busy as NYC. One possible recommendation for this is to force software developers, like Apple, to add more features restricting the phone capabilities when the car is moving. Though there is the obvious hindrance of being potentially cut off from communication, limiting smartphone functionality while driving will foster a safer driving environment with less distractions putting other drivers at risk. However, since there will be a limit to how many restrictions cell phones can have before being considered an unethical breach of privacy, an underlying recommendation would be to raise the penalties of distracted driving, in which additional punishments coming in the form of fines and possible license suspensions would be put in place. Doing so will likely further discourage instances of distracted driving.
Speeding is the Most Deadly Collision Cause
While driver distraction leads to the most collisions, it is both difficult to control for, as distraction is an inevitable part of human driving, and not as deadly on a per-collision basis. Our analysis showed that “unsafe speed” is the leading cause for collisions resulting in fatality. With such a densely populated metropolitan area as New York City, reaching a speed that is considered “unsafe” is very easy. Seeing as how the streets are always incredibly busy with automobiles, pedestrians, and cyclists, there is usually little room for maneuverability, thus maintaining a slow driving speed is even more important than other cities. Our recommendation is to implement more speed control on the streets, whether it be through physical speed reducers (speed bumps) or to make revisions to speed limits in particularly problematic areas. In 2020, New York’s Mayor, Bill de Blasio, lowered speed limits in certain areas by 5mph to 25mph. The city should consider further reductions only in areas where collisions caused by speeding are occurring at a high rate. The following displays a map showing where collisions labeled as “unsafe speed” are clustering:
Manhattan contains very few instances of “unsafe speed” collisions, suggesting that it is not an area to target with a speed limit reduction. In contrast, there are numerous zip codes throughout Brooklyn and the Bronx worth targeting; it is worth noting that three zip codes in Brooklyn have at least 61 collisions in the sample as a result of unsafe speed, whereas Manhattan does not even contain a zip code with more than 20.
Using Mapping to Target Certain Areas
On the topic of targeting problematic areas, we have mapped out certain metrics by zip code in order to discern where the city should focus policy. Effective policy cannot be based solely on large-scale solutions; rather, it can be seen that specific areas are clusters for certain types of collisions, and deliberate measures to target those problematic areas to explore exactly why these collisions are occurring could be the most effective method of increasing traffic safety. Simply labelling the cause of the variable (i.e. driver inattention/distraction) does not nearly enough information; there is very likely a larger story to tell about each collision. The map below displays a heatmap by New York zip code of where collisions are clustering.
It is clear that collisions cluster very strongly on the east side of Brooklyn and the Northern Bronx. The city can take this data and explore into these zip codes with a high number of collisions on the ground-level and figure out on a deeper level why these collisions are occurring.
Collisions are Far More Common at Night
From our analysis, it is easy to see that collisions occur at a much higher rate between the hours of 5pm to 7am, which coincide with both sunset and sunrise respectively. We attributed this relationship to the issue of visibility, seeing how accidents begin to rise as New York gets darker and gradually continues to increase until the morning, shown below.
With navigation itself already being difficult at night, insufficient lighting also magnifies the effect distractions have on drivers, which then hinders their reaction times. As a result, drivers may not notice important road signs or even other vehicles until it is too late to react. In addition to visibility, roads at night will also be populated with drivers that create a hazardous driving environment, mainly due to either drowsiness or being under the influence of alcohol. In turn, tired and drunk drivers pose the risk of poor or unpredictable driving behavior which further increases the frequency of collisions. Our recommendations to combat the difficulties of night time driving are to enhance the lighting and visibility on roads and to the lower the BAC limit from .08 to .05. By either installing more street lights to ensure sufficient lighting on roads and bringing more attention to important road signs, drivers will have an easier time seeing and reacting to any potential hazards. In addition, lowering the BAC will drastically reduce the amount of driving under the influence on the road which would reduce the hazardous driving conditions that out of other drivers’ control. Since this is already being discussed and seeing favorable opinions on implementation, lowering the BAC limit is a very realistic outcome.
Maybe Humans Just Aren’t Great Drivers
While there are numerous factors to point out that explain how and why collisions are occurring, it became clear as we dug into the data that the majority of collisions are caused by human error. Other than increasing penalties for such errors and other minute policy adjustments, it is difficult to account for the fallibility of humans; while more severe penalties may lead to slightly more conscious awareness of how safe one is while driving, it is unlikely to make a change massive enough to reduce collisions. Our main recommendation to respond to the insight that the vast majority of collisions are due to avoidable human error is to expedite the process of transitioning transportation in the city to more autonomous vehicles. On the West Coast, such vehicles have been tested for years. New York would be best served in the long term beginning to test these vehicles in the city in order to assure the seamless transition when autonomous vehicles are commonplace, especially given the unique layout and difficult driving conditions in New York. One possible way the government could go about starting this process is to incentivize New Yorkers with a tax incentive to invest in vehicles with a certain level of autonomous technology in them, similar to how there was a national tax incentive for getting electric cars a couple years ago. But, of course, it must be ensured that these vehicles are safe in the New York City streets.
Interactive
We developed an interactive app through R Shiny that will allow you explore various causes and their impacts on NYC collisions. The interactive has different tabs as well to better visualize how motorists, pedestrians, and cyclists are impacted differently.
Feel free to explore here.