Repository for advanced traffic forecasting models integrating GCN, LSTM/Bi-LSTM, and attention mechanisms for improved accuracy, including weather data processing.

An eXplainable AI system to elucidate short-term speed forecasts in traffic networks obtained by Spatio-Temporal Graph Neural Networks.

Calculating the nearest weather sensor for each traffic sensor and then merging the weather sensors' temporal data with the traffic sensors'.

Smart Traffic — reproducible multimodal ITS pipeline (public datasets): Kafka streams → Spark Structured Streaming + Delta Lake, MinIO frames, YOLOv8+ByteTrack + LightGBM→ONNX workers, Streamlit UI, Prometheus/Grafana. Docker-first, CPU-only, Windows-friendly.

Robust spatiotemporal traffic forecasting on METR-LA & PEMS-BAY with GRU + graph-temporal models, strict time splits, robustness tests, and conformal uncertainty.