How to install

How to install#

Tip

To speed-up installation, replace conda with mamba (see Mamba docs), or replace pip with uv pip (see Uv docs).

conda install -c conda-forge geoutils

Or with Pip:

pip install geoutils

Use "geoutils[opt]" with Pip to include all optional features, or select the extras listed below. Extras can be combined, for example:

pip install "geoutils[io,paral]"

Caution

If a pip update breaks your installation, make a fresh install in a new environment with your other packages.

git clone https://github.com/GlacioHack/geoutils.git
cd geoutils
conda env create -f dev-environment.yml
conda activate geoutils-dev

These main dependencies themselves depend notably on NumPy, pandas, PyProj and Pyogrio.

Optional dependencies are grouped in the following pip extras (opt includes all of them):

gstat: SciKit-GStat for variogram estimation and fitting, and GSTools and GPyTorch (with PyTorch) for kriging and random fields.
cloud: Fsspec with HTTP and S3 support for remote file access.
io: LasPy with Lazrs for LAS/LAZ/COPC files, PyArrow with GeoPandas 1.1 and above for GeoParquet, and netCDF4 and Zarr for array storage.
paral: Dask, Distributed and Dask-GeoPandas for out-of-memory and distributed parallel execution.
num: Numba for faster filters and point cloud gridding, Flox for grouped statistics, and ExactExtract for fractional polygon coverage.
plot: Matplotlib for raster, vector and statistical plots.
prof: Psutil and Plotly for profiling execution time and memory.