wader444/road-accident-hotspot-detection ? reverse-engineered prompt

Build me a Python project that finds road accident hotspots from latitude and longitude data and shows them on an interactive map. I want it to clean the accident coordinates, filter them to the Miami area, cluster the points with DBSCAN using haversine distance, and also compare the results with K means so I can see the difference between density based grouping and fixed cluster grouping.

Please generate the analysis outputs too, like a summary of the biggest hotspots, a bar chart of the top clusters, and a Folium map I can open in a browser with accident points and hotspot centers. It should also be able to prepare a different city from a large raw accidents file without loading everything into memory at once, then run the full pipeline on that city. If you need to check anything, look up current docs online. I also want clear scripts I can run step by step from the command line and sensible default settings so it works out of the box.