lukmani666/nigeria-food-market-price-tracker
Domain:
agriculturesocioeconomicRecord type:
software
Creator:
luk
Host:
# 🛒 Smart Price Tracker for Food Markets (Nigeria)
A production-ready machine learning application for predicting food prices in Nigerian markets using XGBoost and LSTM models.
## 🌟 Features
- **🎯 Price Prediction**: Accurate food price predictions using state-of-the-art ML models
- **🗺️ Market Finder**: Discover markets with the lowest prices for specific commodities
- **📊 Data Explorer**: Interactive visualization and analysis of market trends
- **🔄 Model Comparison**: Compare XGBoost and LSTM performance metrics
- **📍 Geographic Visualization**: Interactive maps showing market locations
- **📈 Historical Analysis**: View price trends and patterns over time
## 🚀 Quick Start
### 1. Install Dependencies
```bash
pip install -r requirements.txt
```
### 2. Train Models
**Option A: Using Python Script (Recommended for production)**
```bash
python run_training.py
```
**Option B: Using Jupyter Notebook (Recommended for exploration)**
```bash
jupyter notebook model_training.ipynb
```
### 3. Test Models (Optional)
```bash
python test_prediction.py
```
### 4. Run Application
```bash
streamlit run app.py
```
The app will open at `localhost`
## 📊 Models
### XGBoost Pipeline
- Gradient Boosting with StandardScaler
- LabelEncoder for categorical features
- Optimized hyperparameters
- Fast inference time
### LSTM Neural Network
- 2-layer LSTM architecture
- Dropout regularization
- Early stopping
- Captures temporal patterns
## 📁 Project Structure
```
├── data/
│ └── nigeria_food_dataset.parquet # Dataset (54,883 records)
├── models/ # Trained models (created after training)
├── model_training.ipynb # Interactive training notebook
├── run_training.py # Automated training script
├── test_prediction.py # Model testing script
├── app.py # Streamlit web application
├── requirements.txt # Python dependencies
├── P …