Analysis and prediction of crimes in New York City using a machine learning approach and spatiotemporal data :
The solution we propose :
- allows predicting a crime with 0,55 as acuuracy, based on provided user data.
- displays the spatial distribution and crime risk areas.
The data originates from the collection of positions and information relating to crimes by citizens and authorities.
We went through different steps and made use of different techniques to realize this solution :
- Explore and clean data
- Model selection and testing
- Web interface development