I have year-long 2024 taxi-trip records for New York, San Francisco, and Chicago that come from fully public datasets. I want you to turn these raw logs into a clear story of when and where people actually travel in each city. The core task is to cluster pickup and drop-off coordinates with DBSCAN, then dig into the spatio-temporal patterns that emerge. I’m especially interested in: • pinpointing the true peak hours and hot-spot locations, and • showing how those peaks shift across the four quarters of each day and as the seasons change through the year. You can work in Python using libraries such as pandas, scikit-learn, GeoPandas, PySAL or any mapping stack you prefer (Folium, Kepler.gl, Plotly, PostGIS, etc.). Feel free to propose an alternative toolchain if it improves performance or visual clarity—just keep DBSCAN at the heart of the clustering step. Deliverables 1. Fully commented code or notebooks that load, clean, and geospatially index the public datasets. 2. A reproducible DBSCAN workflow with parameter-selection logic explained. 3. Interactive or high-resolution static maps that highlight cluster centers and their evolution over time. 4. A concise report (PDF or Markdown) interpreting the findings, with separate sections for each city and cross-city comparisons. 5. A README describing setup steps so I can rerun everything locally. Acceptance criteria • Clusters must be evaluated for spatial coherence and temporal stability. • Peak-hour definitions need quantitative backing (e.g., trip count thresholds). • Visuals should allow me to see at a glance how patterns differ morning vs. evening and winter vs. summer. • All deliverables must execute start-to-finish on a fresh environment using only public data sources. If this sounds straightforward for you, let’s get started—I’m ready to review a brief outline of your approach and timeline.