Wednesday, May 16, 2018

The Car Crash Prediction System


A new mapped based tool claims to be able to predict the likelihood of automobile collisions for every street and intersection. It uses machine learning and historical collision data to determine the future risk of car crashes on city streets.

The Louisville Automobile Collision Prediction System color-codes sections of Louisville's roads based on their predicted collision risk score. Using the colors of the roads you can determine the predicted risk of a future car collision. If you select individual road segments on the map you can view the underlying historical collision data on which the collision risk score is based.

The slide control in the map control panel allows you to filter the section and intersections shown on the map by a risk score range. This therefore allows you to search the map for the most risky sections or intersections or to locate the safest roads and intersection in Louisville.

You can read more about how the collision prediction system works by clicking on the 'Full Report' link above the map.

If you remain a little skeptical about these types of precog maps then you might enjoy the satirical White Collar Crime Risk Zones interactive map.
Post a Comment