Crime Forecasting to Enhance Safety Measures in São Paulo Using Machine Learning

Local Chapter São Paulo, Brazil Chapter

Coordinated byBrazil ,

Leonardo Yuiti Ono

Status: Ongoing

Project background.

Public safety is a topic of great importance, and predicting crimes can be a valuable tool for resource allocation and strategic planning for law enforcement agencies. In this project, we aim to develop a machine learning model to predict the number of crimes by crime category, month/year and region inside of São Paulo State. Additionally, we will create an interactive web app to visualize crime predictions on a map, allowing users to explore the predictions for future dates.

The problem.

Predicting future crimes can provide valuable insights for decision-making and the planning of public safety policies. With a reliable prediction model and an interactive web app, we can empower users to visualize crime predictions by region in an intuitive format, facilitating the understanding and utilization of this information to enhance public safety.

Project goals.

- Develop a machine learning model capable of predicting the number of crimes by region and crime category for future periods, using a historical dataset containing information about crimes. - Create an interactive web app that allows users to visualize crime predictions on a map, with the ability to filter by region, crime category, and specific dates. - Facilitate the understanding and utilization of future crime predictions, making the information accessible and useful to the general public.

Project plan.

  • Week 1

    Onboarding; literature review; code availability; brainstorming; problem statement.

  • Week 2

    Data collection and preparation.

  • Week 3

    Exploratory data analysis.

  • Week 4

    Development of a predictive model.

  • Week 5

    Testing and validation.

  • Week 6

    Web app development.

  • Week 7

    Web app deployment; project review.

Share project on: