If It Fitz: Tracking Illegal Parking with Machine Learning

A sight too common in NYC: a car parked on a bike lane and a cyclist forced onto traffic. For cyclists this situation can lead to dangerous environments where cyclists have lost their lives. As a cyclist myself I’ve found myself walking my bike to avoid this same scenario. But how do we quantify this problem? How can we know this is an extensive problem?

Thankfully computer scientist Alex Bell was able to train a machine learning (ML) algorithm using camera footage to track cars parked on bike and bus lanes with better accuracy. Before, the best data we had on illegal parking were reports from NYC’s 311 app, which although a good start they only capture a small percentage of illegal parking instances. By running this ML algorithm on traffic cameras we can gain a better understanding of what streets are hot beds for illegal parking. This information can help us better deploy police officers to areas with high traffic violations to mitigate their danger while we deploy better bike lanes.

However as the picture above shows, authorities and other public servants will need to meet us halfway if we are to make our cities safer for everyone.