lunaiile/air-quality-predictor

Air quality PM2.5 predictor for Algiers using Weka M5P # Air Quality Predictor — Algiers 🌍 A desktop Java application that predicts **tomorrow's PM2.5 air quality** for Algiers, Algeria using machine learning (Weka M5P model tree). Built as a university time-series analysis project. --- ## Features - **AI prediction** — M5P model tree trained on historical PM2.5 data, predicts next day's value - **Intuitive input** — select how the air feels today (Fresh / Clean / Hazy / Heavy / Thick) as the current-day signal - **Live clock** — real-time date and time displayed in the header - **Historical stats** — shows average, worst, and best recorded PM2.5 from your dataset - **AQI result card** — color-coded prediction with health advice and affected groups - **Professional UI** — custom-painted Swing components, no emojis, no third-party UI libraries --- ## Project Structure ``` airquality/ ├── src/ │ └── timeseries/ │ ├── Main.java # Entry point │ ├── MainWindow.java # Full UI (Swing) │ ├── AirQualityRecord.java # Data model (date + PM2.5) │ ├── AirQualityDataLoader.java # CSV parser │ ├── FeatureEngineer.java # Feature extraction for ML │ └── WekaModel.java # Weka classifier wrapper ├── data/ │ └── algiers_sample.csv # Sample dataset (replace with your own) ├── lib/ │ └── weka.jar # Weka library (add manually — see Setup) └── README.md ``` --- ## Setup ### Requirements - Java 17 or higher - Weka 3.8 — download `weka.jar` ### Steps 1. **Clone the repository** ```bash git clone github.com cd air-quality-predictor ``` 2. **Add Weka** - Download `weka.jar` from waikato.github.io - Place it in the `lib/` folder 3. **Compile** ```bash mkdir -p out javac -cp lib/weka.jar -d out src/timeseries/*.java ``` 4. **Run** ```bash java -cp out:lib/weka.jar timeseries.Main ``` On Windows use `;` instead of `:` in the classp …