Geospatial Data with Open Source Tools in Python

This tutorial will focus on open source libraries that provide a high-level, pythonic interface to geographic data and computations. Students will learn to read standard GIS file formats, perform spatial calculations, and plot results.

Tutorial materials

• Presentation slides
• Examples
• Exercises

Installation

The following packages will be required for this tutorial:

• pyproj
• Fiona
• Shapely
• rasterio
• geopandas

Optional packages that may be used in demonstrations, but not required for exercises, include:

• basemap
• Cartopy
• geojsonio.py

These packages have a number of prerequisites, including NumPy, pandas, matplotlib, and GDAL. I recommend starting with a standard scientific python distribution such as Canopy or Anaconda.

All packages are supported for Python 2.7 and recent versions of Python 3 (3.4 is recommended). Any of the 3 major platforms (Windows, OS X, and Linux) should work. Particularly on Windows, using precompiled packages when available will usually be much easier and less error-prone.

Installing in Canopy

Fiona, Shapely, and pyproj (as well as optional packages GDAL, basemap and cartopy) can be installed with the Canopy package manager, or from the command line using enpkg. Then follow the instructions for installing rasterio and geopandas with pip below.

Installing with conda

Many of the packages are available in Anaconda. Matt Craig contributed the script conda-setup.sh which installs all of the required packages (and most of the optional ones) for this tutorial.

Installing with pip

First, make sure you are using the most recent version of pip available, either by updating from your distribution's package manager or by using pip install --upgrade pip.

For most of the packages, pip install <packagename> will be sufficient. Please install the most recent development version of geopandas with the command:

pip install git+git://github.com/geopandas/geopandas.git

Alternatively, you can install the source package from the geopandas GitHub repository and install it into your python if you are comfortable doing so.

Testing your installation

Run the check_env.py script. If all the required packages say [ OK ], you should be ready.

Data

The data for this tutorial can be download at this location. Please download and unzip this file in this directory. Alternatively, run the download_data.py script to handle the download and unpacking for you.