The goal of my project is to build a machine learning that predicts the number of traffic injuries at a given intersection of NYC and use the results to suggest the most cost-effective improvements to reduce injuries, while maintaining efficient traffic flow..

This project will a large impact by identifying mechanisms for change to enable cities' transportation and mobility to interact safely with its citizens.

I will generate features from a variety of datasets that can be separated into two general categories - First, features of the streets such as the number of driving lanes, number of bike lanes, etc, as predictors of the number of injuries. Second, inferring the pedestrian density of the area as a predictor of the number of injuries. The pedestrian density may be inferred from the economic activity such as number of tourist attractions or proximity to subway stations.

By understanding how the various features interact we can leverage the most impactful features to obtain an optimal solution that decreases traffic-related injuries and maintains traffic flow.

Alternate applications from this project:

• Estimating pedestrian, car, or cyclist density in a city for increasing advertising potential.

• Determining where to place a food truck to maximize profitability.

• Location-based optimization of car/home/health insurance premiums.