Doing Geospatial in Python

With a low barrier to entry and large ecosystem of tools and libraries, Python is the lingua franca for geospatial development. Whether you are doing data acquisition, processing, publishing, integration or analysis, there is no shortage of solid Python tools to assist in your daily workflows.

This workshop will provide an introduction to performing common GIS/geospatial tasks using Python geospatial tools such as OWSLib, Shapely, Fiona/Rasterio, and common geospatial libraries like GDAL, PROJ, pycsw, as well as other tools from the geopython toolchain. Manipulate vector/raster data using Shapely, Fiona and Rasterio. Publish data and metadata to OGC web services using pygeoapi, pygeometa, pycsw, and more. Visualize your data on a map using Jupyter and Folium. Plus a few extras in between!

The workshop is provided using the Jupyter Notebook environment with Python 3.


The workshop uses Jupyter Notebooks. Jupyter is an interactive development environment suitable for documenting and reproducing workflows using live code.

As the installation of all dependencies on all platforms (Windows, Mac, Linux) can be quite involved and complex this workshop provides all components within a Docker Image.

In addition, geospatial web services like pygeoapi and pycsw in this workshop are provided by Docker images.

The core requirement is to have Docker and Docker Compose installed on the system. Once you have Docker and Docker Compose installed you will be able to install the workshop without any other dependencies.

Optional requirements

Users may optionally install QGIS as a GIS data viewer. QGIS is a multiplatform open source desktop application for geospatial data.


The workshop provides various sample data to illustrate Python geospatial functionality which has been tested to cover the workshop requirements.

Having said this, please feel free to bring your own! Examples:

Verifying your environment

Ensure Docker is running on your computer, then verify that the docker and docker-compose commands are working and available:

docker version
docker-compose version


curl -O
unzip master
cd geopython-workshop-master/workshop
# start the workshop
./ start
# check the URL+token
./ url

# open browser to resulting URL+token
# or if on Linux/MacOS, run:
# ./ url | xargs open

# stop workshop
./ stop

If the above .sh scripts do not work on your system you can execute docker-compose directly via:

# in dir geopython-workshop-master/workshop
docker-compose up -d
docker logs geopython-workshop-jupyter
# look for URL+Token and Copy/Paste in browser

Installation Issues

Docker installed but problems installing/running the workshop? Below some tips:

Download Problems

Although curl may be on your system it may have problems with SSL (one user noted using OSGeo4W). In that case you can copy/paste the download URL in your browser and download from there.

File/Drive Sharing

The workshop setup involves Docker Volume Mounting. For Mac OS and Windows installs be sure to enable File/Drive Sharing within Docker Desktop for the directory where you unzipped the workshop. Go to the Preferences/Settings | File Sharing... menu and make settings accordingly.

Running in VirtualBox

You may also run a VirtualBox VM with preferably Ubuntu, install Docker there and run the workshop. Even better if you use Vagrant to provision/manage your VM. You could even unpack the .zip file on your local machine and mount it within the VM, start the workshop there.

In any case, in order to access the services from your local machine, you need to do port mapping from ports within the VM to your local machine in order to access the workshop from your local browser. The following ports need to be mapped from the VirtualBox VM to your local system: 8888 (Jupyter), 5000 (pygeoapi) and 8001 (pycsw) .

You will possibly need to enable firewall access for these ports within your VM. Do this as follows:

sudo ufw allow 8888/tcp
sudo ufw allow 5000/tcp
sudo ufw allow 8001/tcp

Within VirtualBox menu you can then map these ports to the same ports on your local system, so the workshop is accessed with your local browser via, etc.

Cannot Access URL

The workshop should run on"token" but in some cases this may not work. In that case you could also try"token".

No Docker Installed?

If you somehow were not able to install Docker: there is a Cloud version of the Jupyter-Notebook-part of the workshop via "Jupyter Binder".

With some limits (e.g. no local geo-services, no data publication), you may follow most of the workshop using a remote Docker instance within your browser via "Jupyter Binder". Click on the button below to launch the Workshop Binder Instance. Startup takes a while, be patient...


Additional notes for Binder session:


A Gitter channel exists for discussion and live support from the developers of the workshop.

Bugs and Issues

All bugs, enhancements and issues can be reported on GitHub.