Malaria Outbreak Prediction Model

A machine learning system that predicts malaria outbreaks in Cameroon using environmental, healthcare, and socioeconomic factors through Random Forest classification.

Project Structure

MALARIA/
├── .scannerwork/
├── venv/
├── .coverage
├── cleaning.py                # Data cleaning & feature preprocessing
├── confusion_matrix.png       # Saved confusion matrix from model training
├── coverage.xml
├── database_schema.sql        # MySQL schema for all dim/fact tables
├── dim_dates.csv
├── dim_demographics.csv
├── dim_environment.csv
├── dim_health_initiatives.csv
├── dim_healthcare.csv
├── dim_infrastructure.csv
├── dim_location.csv
├── dim_prevention.csv
├── dim_socioeconomic.csv
├── dim_weather.csv
├── fact_malaria_cases.csv
├── feature_importance.csv     # Exported feature importance scores
├── feature_importance.png     # Feature importance plot
├── loading.py                 # ETL: load all CSVs into MySQL
├── malaria_prediction_model.py# Train & evaluate Random Forest model
├── predict_malaria.py         # CLI tool for interactive risk prediction
├── processed_data.csv         # Fully processed dataset (optional artifact)
├── requirements.txt           # Python dependencies
└── test_malaria_db.py         # Pytest suite for ETL & preprocessing

Overview

This project implements a predictive model for malaria outbreak detection using machine learning techniques. The system analyzes multiple dimensions of data including:

• Environmental factors
• Demographic information
• Healthcare accessibility
• Infrastructure data
• Prevention measures
• Socioeconomic indicators
• Weather patterns

Data Sources

The project uses a dimensional data model with the following key files:

• Fact Table: fact_malaria_cases.csv - Contains the main malaria case records
• Dimension Tables:
  • dim_dates.csv - Temporal dimensions
  • dim_demographics.csv - Population demographics
  • dim_environment.csv - Environmental factors
  • dim_health_initiatives.csv - Health programs and initiatives
  • dim_healthcare.csv - Healthcare facility information
  • dim_infrastructure.csv - Infrastructure availability
  • dim_location.csv - Geographical information
  • dim_prevention.csv - Malaria prevention measures
  • dim_socioeconomic.csv - Socioeconomic indicators
  • dim_weather.csv - Weather-related data

Key Components

Data Processing

• cleaning.py: Handles data cleaning, table joins and feature preprocessing
• loading.py: Loads all CSV source files into the MySQL malaria database

Model

• malaria_prediction_model.py: Trains and evaluates the Random Forest outbreak classifier
• predict_malaria.py: Interactive CLI that uses the trained model to estimate outbreak risk
• test_malaria_db.py: Pytest suite for ETL and preprocessing logic

Analysis Outputs

• confusion_matrix.png: Visual representation of model performance
• feature_importance.csv: Detailed feature importance scores
• feature_importance.png: Visual representation of feature importance
• processed_data.csv: Final processed dataset used for modeling

Installation

1. Clone the repository
2. Create a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install required Python dependencies:

pip install -r requirements.txt

4. Install and configure MySQL:

• Install MySQL Server (5.7+ or 8.x)
• Ensure the server is running on 127.0.0.1
• By default, the scripts expect:
  • user: root
  • password: your root password
  • database: malaria
• If your credentials differ, update the connection settings in:
  • cleaning.py
  • loading.py
  • malaria_prediction_model.py

5. Create the malaria schema and tables in MySQL:

• Open database_schema.sql in MySQL Workbench (or another client)
• Execute the entire script to create all dim_* tables and fact_malaria_cases

Usage

1. Load all CSV data into MySQL:

python loading.py

2. (Optional) Build a fully processed CSV for offline analysis:

python cleaning.py

3. Train and evaluate the Random Forest model:

python malaria_prediction_model.py

This will:

• Simulate 2 years of historical data
• Train a Random Forest with class balancing (SMOTE) and GridSearchCV
• Save the best model to malaria_model.pkl
• Generate updated feature_importance.png and confusion_matrix.png

4. Run the interactive prediction CLI (after malaria_model.pkl exists):

python predict_malaria.py

You will be prompted for environmental, healthcare and socioeconomic inputs, and the tool will display an outbreak risk level and recommended actions.

5. Run Tests:

python -m pytest test_malaria_db.py

Model Evaluation

Model performance metrics and visualizations can be found in:

• confusion_matrix.png - For model accuracy evaluation
• feature_importance.png - For understanding feature significance

Contributing

1. Fork the repository
2. Create your feature branch
3. Commit your changes
4. Push to the branch
5. Create a new Pull Request

License

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

---

Note: This project uses dimensional modeling for data organization. Make sure all dimension tables are properly loaded before running the fact table integration.