Installation¶
From PyPI¶
1 | |
Requires Python 3.11 or newer. The dependencies (geopandas, pandas, matplotlib, pyarrow, requests) are installed for you. geopandas is the heaviest; if pip struggles with it on your system, installing it first with conda (conda install geopandas) is the usual fix.
Check it worked:
1 | |
From a clone of the repository¶
Clone the repository if you want the scripts that build the full history, the Explorer code, or the committed master dataset. They are not part of the PyPI package.
1 2 3 | |
What you get either way¶
pip install |
git clone |
|
|---|---|---|
The geocrime_stl library (download, clean, export, summarize) |
yes | yes |
| City, ward and neighborhood boundary files used by the library | yes | yes |
run_pipeline.py (add one month) and download_historical_data.py (build the full history) |
no | yes |
The master dataset (data/processed/slmpd_crime_data.parquet) |
no | yes |
The Explorer source (dashboard/) |
no | yes |
Generated data goes into a data/ folder. In a clone it sits at the repository root. With a plain pip install it is created in the folder you run Python from. Set the environment variable GEOCRIME_DATA_DIR to put it somewhere else.