Skip to main content

pinsheff

16 GIS with Python

The Council’s ArcGIS software from ESRI has a much stronger integration with Python than R, which is why this “GIS with Python” chapter precedes the “GIS with R” chapter.

There are three main ways of leveraging GIS (in particular the Council’s ArcGIS) via Python. In increasing order of difficulty in terms of initial setup:

  1. ArcGIS Notebooks (see the sub-heading immediately below)

  2. Non-ESRI IDE e.g. Jupyter Notebooks, VS Code or RStudio

  3. R via the reticulate package

ArcGIS Notebooks take care of most the Python environment and ESRI license settings. Taking your Python scripts outside of ArcGIS Notebooks, e.g. by using VS Code, means that you’re taking more responsibility for the Python environment and ESRI license settings. Using Python in R via the reticulate package means that you have the added responsibility of the R environment.

16.1 ArcGIS Notebooks

ArcGIS Notebooks are built on top of Jupyter Notebook, an open-source web application that allows you to create and share documents that contain live Python code, visualizations, and narrative text.

ArcGIS Notebooks can be created in ArcGIS Pro and AGOL. They are also now available in ArcGIS Enterprise but have yet to be setup in the Council’s Portal.

Learn Python with ArcGIS Notebooks

Notebooks in ArcGIS Pro documentation

ArcGIS API for Python Fundamentals learning plan

16.2 ArcGIS Python API

The arcgis package is used to work with ArcGIS Enterprise (Portal) and AGOL.

- The name of the package and it’s purpose - OneNote?

To use the arcpy or arcgis packages we first need to specify the ArcGIS Python environment

16.3 ArcPy

The arcpy package is used to leverage the ArcGIS Desktop or Pro functionality. Like the arcgis package, to use the arcpy package we first need to specify the ArcGIS Python environment.

The distinction between web and desktop, for arcgis and arcpy packages respectively, is a bit of a simplification but it serves a purpose. It’s worth noting that the integration between the two packages isn’t as seamless as you might expect.

TODO

- setup

- cross ref Docker

16.4 Non-ESRI GIS with Python