For most drivers, speeding is a rare mistake. For some, it’s a dangerous habit. But for everyone on our roads, it’s a leading cause of death.
In New York City, speeding contributes to roughly one in three traffic deaths, with crashes killing more than 1,000 New Yorkers across the state in 2024 alone. These aren’t just numbers; they represent mothers, brothers, friends and futures cut short.
We know speeding is preventable — and as college freshmen at the Stevens Institute for Technology and rising data scientists and software engineers, we wanted to find a way for our skills to help prevent tragedy. We’ve heard stories of those whose lives were forever changed by traffic crashes, we’ve seen people who have been harmed by speeding and we’re afraid for our friends and family, but this is an issue that’s easily preventable with modern technology.
In December, we joined — and won — a "hack-a-thon" dedicated to using data in the fight for safe streets, organized by Data Science for Social Good, BetaNYC, Families for Safe Streets and Transportation Alternatives. We wanted to answer a critical question: how do we slow down the worst-of-the-worst repeat speeders? The Stop Super Speeders bill (S.4045/A.2299) requires the most reckless super speeders — those with 16 or more speed safety camera tickets, or who have accrued 11 or more points on their license related to speed — to install speed limiters to prevent their vehicles traveling more than 5 mph above the speed limit. It’s similar to how those who drive drunk are required to install breathalyzers — a proven technological way to save lives on our roads.
There’s a lot of data about super speeders — but it’s dispersed among hundreds of datasets from dozens of places. Looking for the right numbers is more than a needle in a haystack, it’s an earring at the bottom of the ocean.
Luckily, with the right tools, the metal detectors built of code, we can quickly and easily organize data about the worst-of-the-worst repeat speeders in New York State into a clean database.
For the hack-a-thon, we were tasked with finding a way to quickly and easily organize data about the worst-of-the-worst repeat speeders in New York State that was clean and user friendly. We wanted to create a proof of concept: that the potentially confusing data architecture required for setting up a system for identifying and mitigating the danger of Super Speeders could in fact be done. These are core technical issues that will need to be addressed when Stop Super Speeders becomes law, and we wanted to help write the first draft of history.
We built a simple, transparent tool designed specifically for non-technical DMV staff and administrators to identify repeat dangerous speeding behavior. At a high level, the system ingests speeding and traffic violation data, cleans and standardizes it, and applies clear, rule-based thresholds to flag drivers or license plates that meet the definition of a super speeder.
How does it work? The tool takes in violations over a certain period of time and focuses on key data points such as the number, severity and timing of speeding offenses to identify repeat high-risk behavior, and transforms these numbers into clear tables or dashboard-style summaries that quickly show whether someone is flagged as a super speeder and why.
Taking something theoretical and making it tangible marks a big step forward. While today, the tool still relies on manually-provided or batch-uploaded enforcement data rather than real-time feeds, and is intentionally conservative to avoid overreaching, it’s a prototype for the future. Identifying patterns that already exist — but are buried in fragmented systems — the State can efficiently find and slow down drivers before the next crash without overloading agency capacity.
What we created in a single day is a beginning rather than a final product.Any real-world deployment would require careful policy design, agency oversight and ongoing evaluation to ensure accuracy, fairness and public trust. This work should be understood as a foundation — a proof that earlier, data-informed intervention is feasible — not the final answer to traffic violence.
New York already possesses information, knowledge and urgency. Speeding is preventable, not inevitable, and we will stop super speeders. New York State can lead the way by passing the Stop Super Speeders Act, knowing there is a community of current and aspiring data scientists who are eager to create the most life-saving program possible.
