Installation

Conda

The foolproof way to install OSMnx is with conda or mamba:

conda create --strict-channel-priority -c conda-forge -n ox osmnx

This creates a new conda environment and installs OSMnx into it, via the conda-forge channel. If you want other packages, such as jupyterlab, installed in this environment as well, just add their names after osmnx above. To upgrade OSMnx to a newer release, remove the conda environment you created and then create a new one again, as above. See the conda-forge documentation for details.

Docker

You can run OSMnx + JupyterLab directly from the official OSMnx Docker image.

Pip

You can also install OSMnx with uv or pip (into a virtual environment):

pip install osmnx

OSMnx is written in pure Python and distributed on PyPI. Its installation alone is thus trivially simple if you have its dependencies installed and tested on your system. However, OSMnx depends on other packages that in turn depend on compiled C/C++ libraries, which may present some challenges depending on your specific system’s configuration. If precompiled binaries are not available for your system, you may need to compile and configure those dependencies by following their installation instructions. So, if you’re not sure what you’re doing, just follow the conda instructions above to avoid installation problems.