ethanol-cx / heterogeneous-crime-prediction-tool

A web-based tool that predicts crimes based on heterogeneous spatial patterns

Heterogeneous Crime Prediction Visualization Tool

Table of Contents

• Introduction
• Getting Started

Introduction

This is a crime prediction visualization tool that allows users to provide crime data, filter data, and see the process of clustering. This relies on a crime prediction method and an evaluation metric that are proposed by the Data Science Lab at USC. Technologies/packages used: Django, Mapbox GL JS, Spectre.css, jQuery, Keras(tensorflow backend), Scikit-learn. The tool needs to be used when connected to the internet.

Getting Started

Please follow these steps before running the code for the first time.

1. Install Python 3
2. It is strongly recommended to use virtualenv to manage packages for this tool. Please see [virtualenv] for more details. If you would use virtualenv, you would need to activate it before
3. Install required packages that are listed in requirements.txt. If you have pip installed, you could run pip install -r packages.txt to install all the packages.
4. Start the server by python djmaps/manage.py runserver. The application will appear in http://127.0.0.1/index
5. Open up the browser and use the tool! There are few things to notice:
   • Please upload the data you would like to use first. The data has to be formatted in a JSON input file with the following format: [[<Type>, <Latitude>, <Longitude>, <Date: mm/dd/yyyy>]]
   • The sample DPS data is located in: djmaps/maps/templates/DPSUSC.json.
   • All the datapoints provided that are outside of the USC "border area" that we defined will be ignored.