Geospatial Environment Installation Guide

Learning Objectives

  • Understand the role and significance of package managers in the context of software development and geospatial data science.

  • Create and activate a new conda environment.

  • Use conda to install geospatial packages from the conda-forge channel.

  • Recognize key geospatial packages

Review

Geospatial Environment Installation Guide#

In the world of software development, package managers have become a fundamental tool for simplifying the installation of software packages and their dependencies. They automate the process of installing, upgrading, configuring, and removing software packages in a consistent manner. In the field of geospatial data science, where several packages need to be installed, package managers ensure a smooth and efficient setup of your environment.

Follow the instructions to install conda here

Creating and Activating a New Environment#

Each project or workflow may require different versions of packages or even different packages altogether. To manage these variations, it is best practice to create isolated environments for each project. This prevents conflicts between package versions and ensures that your projects remain reproducible.

You can create a new environment in conda using the following command:

conda create --name myenv

Replace myenv with the name of your environment.

To activate this environment, use:

conda activate myenv

Now, any package you install will be installed in this environment, keeping it isolated from other environments.

Installing Geospatial Packages#

Before we install anything, we need to tell conda to use the conda-forge channel, which carries the most up-to-date geospatial packages. If you installed Miniforge this is already the default, but setting it explicitly does no harm:

conda config --add channels conda-forge
conda config --set channel_priority strict

Nearly all of the packages we need are available on conda-forge, so we can install them in a single command. Installing them from one channel — rather than mixing conda and pip — is the approach the GeoPandas developers recommend, since pulling geospatial dependencies from different sources can lead to hard-to-debug import errors.

conda install -c conda-forge geowombat geowombat-ml geowombat-stac pyproj geopandas contextily earthpy census geoplot osmnx pykrige us

One package, pytidycensus (used in the census chapter), is not published on conda-forge, so we install that single package from PyPI into the same environment:

pip install pytidycensus

Note

Prefer pip for everything? You can instead install the whole stack into a virtual environment with pip install "geowombat[ml,stac]" pyproj geopandas contextily earthpy census pytidycensus geoplot osmnx pykrige us — just don’t mix pip and conda for the geospatial packages within one environment. Installing the pure-Python pytidycensus with pip on top of a conda environment, as above, is fine.

Now you have all the packages installed and ready to be used in your new environment!

Verifying the Installation#

Next we need to activate the environment and verify that the packages are installed correctly.

First activate the environment you created earlier:

conda activate myenv

Then enter a Python session by typing:

python

You should now see >>> in your terminal, indicating that you are in a Python interpreter session.

To verify that the packages are installed correctly, you can run the following commands in a Python interpreter:

import geowombat as gw
from geowombat.ml import fit, predict, fit_predict
import pyproj
import geopandas
import contextily
import earthpy
import census

Now you can exit the Python interpreter by typing:

exit()

Note

If you encounter any errors during the import, it means that the package is not installed correctly. You can try reinstalling the package or checking for any dependencies that may be missing.