Introduction to Graphical Information Systems Using Python

This repository contains worked examples in Jupyter notebooks for the 3-day online course Introduction to Python GIS by CSC Finland. The materials are originally drawn from the Automating GIS Processes masters course at the University of Helsinki.

Installing GIS Python Modules on Ubuntu Linux 18.04

Performing GIS tasks in Python relies on a suite of python modules that can sometimes have complex dependencies. The easiest way to get a working installation is to build a python environment that contains specific module versions drawn from (mostly) the same packaging source. This has the advantage that you can load the environment as needed, so it won't break existing installations. Here I use the Anaconda Python distribution and create a Python 3.6 environment called gis and install default modules and modules from the conda-forge organisation.

Download and install Anaconda:

From the command line do the following:

wget https://repo.anaconda.com/archive/Anaconda3-5.2.0-Linux-x86_64.sh
bash Anaconda3-5.2.0-Linux-x86_64.sh

My preference is to install the distribution in a shared directory /opt/share/anaconda3_GIS and use a command-line alias to make it visible to the system (I have multiple Python installations that I need for different tasks). If you want to do this edit your $HOME/.bash_aliases file to contain:

alias aconGIS="export PYTHONPATH=/opt/share/anaconda3_GIS; export PATH=/opt/share/anaconda3_GIS/bin:/opt/share/anaconda3_GIS/lib/python3.5/site-packages:$PATH;"

If this is going to be your default Python, you can just add the line without the surrounding alias aconGIS=" ... ". Restart your terminal window, execute aconGIS (if necessary) and check you are pointing to the correct Python install using which python.

Install GIS Modules

Download this repository by clicking on "Clone or Download" above, or executing the following on the command line:

git clone https://github.com/crpurcell/IntroPythonGIS.git

We are going to create a new conda environment and install most Python modules using a yml file (which specifies the environment name, modules and version numbers):

cd IntroPythonGIS
conda env create -f gis_18.04.yml

Activate the environment using the command source activate gis. Several packages need to be installed from the conda-forge repositories or via the pip command.

conda install -y -c conda-forge rasterstats
conda install -y -c conda-forge osmnx
pip install cython
pip install geopy
pip install https://github.com/mullenkamp/PyCRS/archive/master.zip

Lastly, the QGIS software is very useful to check that your Python scripts are doing the right thing. Install this from the Ubuntu repositories using the command sudo apt install -y qgis (admin privileges needed).

Running the Python Jupyter Notebooks

Before running the notebooks, download the example datasets and move the data into a sub-directory IntroPythonGIS/Data. Data are available from here, here, here and here.

To run the Jupyter Notebooks, do the following on the command line:

aconGIS              # (if you made an alias in your .bash_aliases file)
source activate gis  # to activate the python environment
jupyter-notebook     # to start the Jupyter notebook server.

Go to your browser and have fun doing GIS analysis in Python!