bsweehoney/air-quality-forecast — reverse-engineered prompt

Build me a Python project that monitors and forecasts air quality for five major cities using live data from AQICN and OpenWeatherMap.

I want it to pull current air quality and weather data, clean it into a usable dataset, create useful features, train an XGBoost model to forecast AQI, evaluate the model with clear metrics like R squared, RMSE, and MAPE, and save charts or dashboard style outputs so I can quickly understand the results.

Please organize it as a simple data pipeline with separate scripts for ingestion, processing, feature creation, training, and evaluation. Include a requirements file, folders for data, models, and outputs, and a README that explains setup, API keys in a .env file, and how to run each step from the terminal.

Use Python 3.11 with pandas, scikit learn, XGBoost, matplotlib, and pyarrow. Keep it easy to run locally, and look up current API docs online if needed.