RandrianantenainaBearisoa/rice_price_predict

This application predict the rice price in Madagascar. # Rice Price Predict This is an End-to-End application that predict the price of rice in Madagascar based on historical data and various factors like the **weather** and the **exchage rate** between the **MGA** and the **USD**. ## Stacks and tools ### Stacks - Python - Jupyter Notebook - FastAPI - Vue.js - TypeScript - Node.js - Sass ### Tools - Docker & Docker Compose - Git & GitHub - MLflow & Minio & PostgresSQL - Nginx - Prometheus - Grafana - Loki - Promtail - Nginx Prometheus Exporter ### Data sources - World Food Programme (WFP) HumData , wfp_food_prices_mdg.csv - World Bank World Bank Data Catalog , global_fuel_price.xlsx - World Bank World Bank Documents , worldbank_commodity_price.xlsx - Yahoo Finance yfinance - Geolocation Data geopy ## 🚀 How to Run Locally ### Prerequisites * Docker and Docker Compose installed ### Steps 1. Clone the repository: ```bash git clone github.com cd rice_price_prediction docker compose -f docker-compose.dev.yml up -d ``` ## Architecture This is a **production-ready end-to-end ML application** following the data engineering and MLOps paradigm. The system ingests data from multiple sources, performs feature engineering, trains models, and serves predictions via a REST API with a web interface. ### System Overview The application is structured as a **3-tier architecture**: ``` ┌─────────────────────────────────────────────────────────────────┐ │ Frontend (Vue.js + TypeScript) │ │ Interactive UI for prediction input/output │ └─────────────────────────────────────────────────────────────────┘ | HTTP/REST | ┌─────────────────────────────────────────────────────────────────┐ │ Backend API (FastAPI + Python) │ │ - Request validation & error handling │ │ - Model inference & post-processing │ │ - JSON logging for observabilit …