I’m sitting on an Excel file that holds roughly 70,000 latitude/longitude pairs. I need this large set consolidated into about 4,000 representative locations, each accompanied by a radius that fully encloses every original point assigned to it. In other words, think of it as transforming dense raw data into a cleaner, cluster-based map: one row per cluster showing • centre_lat • centre_lon • radius (metres or kilometres—let’s decide what works best) The finished dataset must arrive as a CSV, ready to load straight into downstream GIS and analytics tools. I don’t mind whether you employ Python (pandas, scikit-learn, GeoPandas), R, QGIS, PostGIS, or any other spatial clustering approach—as long as the logic is sound, reproducible, and explained briefly so I can rerun the process later if needed. Timing is critical; I’d like the first draft ASAP. Deliverables are: 1. The CSV containing ~4,000 rows with centre coordinates and radius. 2. The script or workflow used to produce it, with brief notes. All 70k original points must fall within the radius of exactly one resulting cluster. Let me know any assumptions (distance units, projection, max radius limits) before you run the full job so we stay aligned.