What it does
Our ML model predicts the general location of crime hotspots on a particular day of the week. This information could be very useful for a police department seeking to optimise their resource utilisation in order to reduce crimes, and quickly respond to ones that do occur.
How We built it
We build a Neural Network through TensorFlow using over 10 years of training data from the city of San Francisco. The data included the location (LatLng) of the crime, the date & time, day of the week, police district. The model takes input as a day and returns the location where the next crime is most likely to occur. The city was divided using a clustering algorithm into 10 mutually exclusive regions. The output of the model is one of these 10 locations.
Challenges We ran into
We did not have enough time to perfect the Neural Network by tweaking various parameters
What's next for Crime Me a River
Build ML models for more cities!
Built With
- google-maps
- javascript
- keras
- tensorflow
Log in or sign up for Devpost to join the conversation.