This repository contains all the collaborative materials used in our project for the Citadel Datathon. Here, you'll find the data analysis notebooks, visualizations, data files, and the code for our web application demo. Our project aims to explore the socio-economic impacts and health outcomes associated with the consumption of ultra-processed foods in the United States.
data: This directory contains all datasets used throughout the project.notebooks: Jupyter notebooks with extensive Exploratory Data Analyses (EDAs), predictive modeling, and statistical tests.plots: Visualizations generated during the analysis, used both within the notebooks and in our final presentation.web app: Source code for the web application demo that visualizes our findings and model predictions.model: All 70+ models trained for different purposesREADME.md: This file, which provides an overview of the repository.
"The predictive models developed in this study, including Random Forest, ARIMA, and Neural Networks, leverage advanced statistical methods and machine learning to forecast the health impacts of ultra-processed food consumption. These models not only predict the future public health scenarios but also assess the effectiveness of potential interventions aimed at improving dietary habits."
To use the notebooks or run the web application:
- Clone this repository to your local machine.
- Ensure you have Python installed, along with the libraries listed in
requirements.txt. - Launch Jupyter Notebook or JupyterLab to open the
.ipynbfiles. - To run the web application, run
app.py
We welcome contributions from other researchers and the public. If you wish to contribute:
- Fork the repository.
- Create a new branch for your feature.
- Commit your changes.
- Push the branch and open a pull request.
This project is open-source under the MIT license.
This work was made possible through the support of Citadel and the effort of all team members involved in the datathon. We thank the organizers for providing the opportunity and resources that contributed to this research. We turned this github into public straight after the deadline of submission.
