Arrest Predictor - Project 04

---

Team

• Harika Peddiraju
• Ulises Domínguez
• Yuting Chien
• Karla Flores

Topic and Description

Topic:

Arrest predictor

Description:

Based on historical crime data of the Chicago city (2001 - 2021), our Classification Machine Learning Models with the Logistic Regression method seeks to be relevant for Police Dispatchers as a tool for deploying resources in a request of presumable crime. The purpose of the model is to predict an arrest given controllable variables.

Model Selection

We decided to work with Logistic Regression. The testing, training, and weighted average were better than other models like Random Forest Classifier or Neural Networks.

Our Project Process

Tools used

• Python and Jupyter Notebook
• Libraries: pandas, numpy, sklearn, pickle, tensorflow
• Flask
• Tableau
• HTML/CSS
• Javascript
• Bootstrap
• D3.js
• Heroku

Deployment

Arrest Predictor

Data source

City of Chicago Data Portal - Police Department - https://data.cityofchicago.org/Public-Safety/Crimes-2001-to-Present/ijzp-q8t2