Chamiln17/Algeria-Weather-Forecaster

# 🇩🇿 Climate Forecasting System for Algeria > Long-term climate forecasting (2024–2040) using an ensemble of ML models, an RL-based model selector, and a RAG-powered conversational interface — built for the Algiers region. --- ## 🏗️ System Architecture The system is organized as a four-stage pipeline: ``` Raw Data (Open-Meteo) │ ▼ ┌──────────────────────────┐ │ DATA PIPELINE │ Preprocessing → Monthly Aggregation → Feature Engineering │ src/preprocessing.py │ Drought indices (SPI, SPEI), anomalies, rolling stats └──────────┬───────────────┘ │ ▼ ┌──────────────────────────┐ │ FORECASTING MODELS │ SARIMA │ LSTM │ Ridge │ Prophet │ Forecasting_Models/ │ Each produces 204-month forecasts (→ 2040) └──────────┬───────────────┘ │ ▼ ┌──────────────────────────┐ │ RL AGENT │ Q-Learning agent trained on 2019–2023 backcasts │ RL Agent/ │ Selects BEST model per month (dual-variable reward) └──────────┬───────────────┘ │ ▼ ┌──────────────────────────┐ │ RAG + UI │ ChromaDB + e5-small-v2 + Groq (Kimi K2) │ src/rag.py + app.py │ Streamlit dashboard with streaming answers └──────────────────────────┘ ``` --- ## ✨ Features | Feature | Description | | -------------------------- | --------------------------------------------------------------------------------------------- | | 📊 **Data Pipeline** | Missing value handling, monthly aggregation, drought indices (SPI, SPEI), anomaly calculation | | 📈 **Trend Analysis** | Mann-Kendall test (with autocorrelation correction), Sen's slope estimation | | 🔬 **Stationarity** | ADF + KPSS tests with automatic differencing order recommendation | | 🔮 **4 Forecasting Models** | SARIMA (auto-tuned), Bidirectional LSTM, Ridge Regression (polynomial), Prophet | | 🤖 **RL Model Sele …