pyplaces is still in development, so if you run into any bugs please report them here
pyplaces is a Python package meant to streamline the usage of large places datasets, notably Overture Maps and Foursquare Open Places. All of the datasets and past releases can be downloaded to a specific address, place or bounding box. Conventionally, users download the data themselves or use a tool such as DuckDB to download what they need, whereas pyplaces is more integrated, faster, and more precise.
This package is available on PyPI as well as conda-forge:
pip install pyplaces
conda install pyplaces
I recommend that you use Anaconda for your Python environments. (I use miniforge and mamba)
To access any places, all you need is a place, address, or bounding box:
from pyplaces.foursquare_open_places import foursquare_places_from_address
hemenway_st_places = foursquare_places_from_address("204 Hemenway Street, Boston, MA")
Additionally, you can:
- Select and filter on columns
- Get data from past releases
- Inspect dataset schemas
- Find relevant category names for retrieving places data
Most, if not all, of the functions follow the same type of workflow.
The documentation is available here in case you run into any issues.
pyplaces is open source and licensed under the MIT license.
This package uses a geocoder that uses OpenStreetMaps's geocoding service Nominatim. Please abide by their usage policy here.
-
more readable schema output
-
reduce time in the pyarrow parquet reading pipeline
-
speed up model transactions- current implementation is inefficient
-
support more geoparquet datasets
-
basic network analysis for transportation data