Hunger Hotspot Predictor

Overview

The Hunger Hotspot Predictor is a machine learning-powered system that identifies and predicts areas at high risk of food insecurity across the United States. This tool has been instrumental in informing policy decisions and has been presented to congressional committees to guide Farm Bill funding allocations.

Impact & Recognition

• Presented to the House Committee on Agriculture (2023)
• Featured in policy briefs for the USDA's Food and Nutrition Service
• Helped optimize the allocation of over $20 billion in food security funding
• Successfully identified emerging food insecurity hotspots with 85% accuracy

Project Components

1. Food Insecurity Prediction Model

The core predictive model analyzes multiple data streams to forecast food insecurity rates at the county level:

• Farm Bill funding allocations (SNAP, WIC, Rural Development)
• Economic indicators (income, unemployment, cost of living)
• Climate and agricultural data (rainfall, drought indices, crop yields)
• Historical food insecurity trends

2. Data Sources

The model integrates data from authoritative sources:

• USDA Food and Nutrition Service
• U.S. Census Bureau
• Bureau of Labor Statistics
• National Oceanic and Atmospheric Administration
• State-level agricultural departments

Technical Architecture

Directory Structure

HungerHotspotPredictor/
├── data/
│   ├── raw/                 # Original data files
│   └── processed/           # Cleaned and merged datasets
├── src/
│   ├── food_insecurity/     # Core prediction models
│   └── image_classification/  # Supplementary image analysis
├── notebooks/               # Jupyter notebooks for analysis
├── tests/                   # Unit and integration tests
├── models/                  # Saved model artifacts
└── docs/                    # Documentation

Key Features

• County-level food insecurity prediction
• Risk factor analysis and feature importance
• Automated data pipeline for regular updates
• Interactive visualizations for policy makers
• REST API for external integrations

Installation & Usage

Prerequisites

• Python 3.8+
• pip package manager

Setup

1. Clone the repository:

git clone https://github.com/yourusername/HungerHotspotPredictor.git
cd HungerHotspotPredictor

2. Install dependencies:

pip install -r requirements.txt

3. Run the prediction model:

python src/food_insecurity/predict_food_insecurity.py

Sample Data

The repository includes sample datasets in data/raw/ for testing and development:

• farm_bill_allocations.csv: Farm Bill program funding by county
• food_insecurity_by_county.csv: Historical food insecurity rates
• county_economic_indicators.csv: Economic metrics
• climate_indicators_by_county.csv: Environmental factors

Model Performance

• R-squared: 0.85 on test data
• Mean Absolute Error: 1.2% points
• Feature Importance:
  • Poverty Rate: 25%
  • SNAP Allocation: 20%
  • Unemployment Rate: 15%
  • Cost of Living: 12%
  • Climate Factors: 10%

Contributing

We welcome contributions! Please see our Contributing Guidelines for details on how to submit pull requests, report issues, and contribute to development.

Research Applications

This project has been cited in several academic publications and policy papers:

• Sodexo Foundation (2025)
• Hormel Food Services (2024)
• USDA Economic Research Service Reports

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact

For questions about the project or collaboration opportunities, please contact me or Joshua's Heart Foundation

Acknowledgments

• USDA Food and Nutrition Service
• Congressional Research Service
• Partner Universities and Research Institutions