Bike-stations-locations-clustering

A sctipt for clustering the bike stations in a city using their location. First, the model is trained using either K-means or Agglomerative Hierarchical Clustering for a user choosen number of cluster. And second, the clsuters are displayed in using an intertactive html map or matplotlib standard figure.
The intertactive map need to be saved on the computer in order to be displayed and requires the package folium. The standard figure can also be saved by pasing a path using the command lien options. New labed json file is also saved.

Using the static data (json files) provided by JCDecaux, you can visulize the clusters of bike stations of any city. The lattitude and longitude of the city are required to use the interactive map (set to Brisbane by default).

See the requirements.txt file for the dependencies.

Getting Started

The script is lunched through the command line:

python bike_station.py -i ./data/Brisbane_CityBike.json

The command’s line arguments are:

  -h, --help    show this help message and exit
  -n N_CLUSTERS, --n_clusters N_CLUSTERS    number of clusters.
  -l {map,figure}, --plot {map,figure}    method of desplaying cluters: - map: interractive html map. - figrue: standard figrue.
  -m {kmeans,ahc}, --model {kmeans,ahc}    model used for clustering the bike stations.
  -p PATH, --path PATH  path where the figure need to be saved.
  -i FILE_PATH, --file_path FILE_PATH  path of json data file.
  -c CITY, --city CITY lattitude and longitude of the city

Visualization

Brisbane city map clustering (interactive map)

python bike_stations.py -n 5 -l map -i ./data/Brisbane_CityBike.json -m kmeans

Brisbane city map clustering (standard figure)

python bike_stations.py -n 5 -l figure -i ./data/Brisbane_CityBike.json -m kmeans

Lyon city map clustering (45.74846 4.84671)

python bike_stations.py -n 5 -l map -i ./data/Lyon.json -m kmeans -c 45.74846 4.84671