Built a production-grade ETL pipeline in Python analyzing the correlation between Calima (Saharan dust) weather events and last-minute hotel bookings in Gran Canaria. The pipeline extracts meteorological data automatically from the AEMET API (Spanish national weather service), transforms it through a multi-stage cleaning process, and loads it into a structured dataset ready for statistical analysis. Key technical decisions: - Fail-fast strategy: the pipeline halts immediately on any data quality violation, preventing silent errors from propagating downstream - Data contracts with pandera: every schema change is caught automatically before it reaches the analysis layer - Dynamic anomaly detection for classifying Calima weather events without hardcoded thresholds - Pearson correlation analysis between weather anomalies and booking demand patterns - Full CI/CD with automated testing via GitHub Actions Deliverable: a reproducible, documented pipeline with a validated dataset and a statistical report on weather-demand correlation. Reviewed and scored by a senior data engineering panel.