Summary and Schedule

Introduction to Geospatial Raster and Vector Data with Python

In this lesson you will learn how to work with geospatial datasets and how to process these with Python. Python is one of the most popular programming languages for data science and analytics, with a large and steadily growing community in the field of Earth and Space Sciences. The lesson is meant for participants with a basic knowledge of Python and it allows them to familiarize with the world of geospatial raster and vector data. If you are unfamiliar with Python, useful resources to get started include the Software Carpentry’s lesson “Programming with Python” and the book “Think Python” by Allen Downey. In the Introduction to Geospatial Raster and Vector Data with Python lesson you will be introduced to a set of tools from the Python ecosystem and learn how these can be used to carry out geospatial data analysis tasks. In particular, you will work with satellite images and open topographical geo-datasets, and learn how these spatial datasets can be accessed, explored, manipulated and visualized using Python.

Case study - Wildfires

As a case study for this lesson we will focus on wildfires. According to the IPCC assessment report, the wildfire seasons are lengthening as a result of changes in temperature and increasing drought conditions. To analyse the impact of these wildfires, we will focus on the wildfire that occurred on the Greek island of Rhodes in the summer of 2023, which had a devastating effect and led to the evacuation of 19.000 people. In this lesson we are going to analyse the effect of this disaster by estimating which built-up areas were affected by these wildfires. Furthermore, we will analyse which vegetation and land-use types have been affected the most by the wildfire in order to get an understanding of which areas are more vulnerable to wildfires. The analysis that we set up provides insights in the effect of the wildfire and generates input for wildfire mitigation strategies.

Note, that the analyses presented in this lesson are developed for educational purposes. Therefore in some occasions the analysis steps have been simplified and assumptions have been made.

The data used in this lesson includes optical satellite images from the Copernicus Sentinel-2 mission and topographical data from OpenStreetMap (OSM). These datasets are real-world open data sets that entail sufficient complexity to teach many aspects of data analysis and management. The datasets have been selected to allow participants to focus on the core ideas and skills being taught while offering the chance to encounter common challenges with geospatial data. Furthermore, we have selected datasets which are available anywhere on Earth.

During this lesson we will setup an analysis pipeline which identifies scorched areas based on bands of satellite images collected after the disaster in July 2023. Next, we will calculate the Normalized Difference Vegetation Index (NDVI) to assess the vegetation cover of the areas before and after the wildfire. To investigate the affected built-up areas and main roads, we will use OSM vector data and compare them with the previously identified scorched areas.

To most effectively use this material, make sure to download the data and follow the software setup instructions before working through the lesson (this especially accounts for learners that follow this lesson in a workshop).

Python libraries used in this lesson

The main python libraries that are used in this lesson are:

The actual schedule may vary slightly depending on the topics and exercises chosen by the instructor.

Data Sets

  1. Create a new directory on your Desktop called geospatial-python.
  2. Within geospatial-python, create a directory called data.
  3. Download the data required for this lesson via this link (678MB).
  4. Unzip the downloaded file and save its content into the just created data directory.

Now you should have the following files in the data directory:

  • sentinel-2 - This is a directory containing multiple bands of Sentinel-2 raster images collected over the island of Rhodes on Aug 27, 2023.
  • dem/rhodes_dem.tif - This is the Digital Elevation Model (DEM) of the island of Rhodes, retrieved from the Copernicus Digital Elevation Model (GLO-30). The original tiles have been cropped and mosaicked for this lesson.
  • gadm/ADM_ADM_3.gpkg - This is the administration boundaries of Rhodes, downloaded from GADM and modified for this lesson.
  • osm/osm_landuse.gpkg and osm/osm_roads.gpkg - They are land-use poylgons and roads polylines of Rhodes, downloaded from Openstreetmaps via Geofabrik and modified for this lesson.

Software Setup

Python is a popular language for scientific computing, and great for general-purpose programming as well. There are many ways to install Python and the required dependencies. In this workshop, we suggest to use uv for its fast and easy installation process.

Discussion

Software Setup using uv

Please follow the instructions below according to your operating system.

Regardless of how you choose to install it, please make sure you install Python version 3.x (e.g., 3.12 is fine). Also, please set up your python environment at least a day in advance of the workshop. If you encounter problems with the installation procedure, ask your workshop organizers via e-mail for assistance so you are ready to go as soon as the workshop begins.

Open a terminal and install uv following the official installation instructions:

SH 

curl -LsSf https://astral.sh/uv/install.sh | sh

Then make sure you are inside the geospatial-python directory you created during the data setup step by doing:

SH 

cd ~/Desktop/geospatial-python

Finally, run the following command to create a virtual environment and install the required dependencies:

SH 

uv venv --python=3.12 && uv pip install -r https://raw.githubusercontent.com/carpentries-incubator/geospatial-python/main/files/requirements.txt

On Windows, first we install uv using PowerShell following the official installation instructions:

POWERSHELL 

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

After the installation, you may see suggestions on the PowerShell terminal like $env:Path = "C:\Users\username\.local\bin;$env:Path" This means you need to manually add the uv executable to your system’s PATH variable. Please run the suggested command in your PowerShell terminal to add uv to your PATH. Otherwise PowerShell will not recognize the uv command in the next step.

Then make sure you are inside the geospatial-python directory you created during the data setup step by doing:

POWERSHELL 

cd \Users\<Username>\Desktop\geospatial-python

And replace the <Username> pattern (including the angle brackets <>) with your Windows username. Finally, run the following command to create a virtual environment and install the required dependencies:

POWERSHELL 

uv venv --python=3.12; if ($LASTEXITCODE -eq 0) { uv pip install -r https://raw.githubusercontent.com/carpentries-incubator/geospatial-python/main/files/requirements.txt}

After the installation, a .venv directory will be created in the current directory, which contains the virtual environment with all the required dependencies.

Testing the installation

In order to follow the lesson, you should launch JupyterLab. Let’s try it now to make sure everything is set up correctly. You should run the following command in your terminal from the geospatial-python directory:

uv run jupyter lab

Once you have launched JupyterLab, create a new Python 3 notebook, type the following code snippet in a cell and press the “Play” button:

PYTHON 

import rioxarray

If all the steps above completed successfully you are ready to follow along with the lesson!

Callout

Alternative: software setup using Anaconda

If you prefer to use Anaconda, you can follow the alternative setup instructions on this page.