- Identify commercial centers using Points of Interest (POI) data
Create clusters of distinct commercial centers or markets using points of interest data of a city.
- Python3
- Jupyter Notebook
Libraries
- Overpy
- GeoPandas
- Folium
- HDBScan
- Scipy
- Shapely
POI data is collected from an open source OpenStreetMap using OverpassTurbo. Alterntively, This can also be done using "Overpy" library.
For clustering, HDBScan is used which extends DBScan clustering algorithm and converts it into a hierarchical clustering algorithm. Since, we are working on GIS point cluster density based clustering algorithm seems to be more suitable than distance based.
The most significant cluster is identified using statistics to find out the cluster with most number of amenities and also intuitively selecting one with variety of commercial amenities.