This is my final project for submission in the advanced geoscripting university course.

• To practice applying best-practices in scientific programmingn
• To reflect on programming habits and improve them.

I chose to do an ecological analysis using data from the GBIF database (https://www.gbif.org/). GBIF provides free and open access to biodiversity data and has an API (https://www.gbif.org/developer/summary). There is also a python client to the API, called pygbif (https://pypi.org/project/pygbif/), maintained by github user sckott.

I split the project in two parts.

1. An exploratory data analysis using Jupyter Notebook, where I experimented and explored the API, some visualisations and small analysis.

2. A python script with the key analyses and functions from the notebook. I included requests of another API here (Eurostat GISCO API). In the python code, I tried to focus on programming style, readability and documentation of the code. Input can be given in configuration files (-json-format). I furthermore tried to avoid or catch some errors which I stumbled upon.

To run the jupyter notebook and the python code, you need python 3.8 and a number of packages:

• jupyter
• pandas
• geopandas
• numpy
• matplotlib
• descartes
• pygbif
• mplleaflet
• pywin32
• pyproj
• geojson
• requests
• pytest

If using Anaconda, the environment can be set up using the environment.yml:

conda env create -f environment.yml

or if you're working with another operating system than windows with the environment_cross-platform.yml:

conda env create -f environment_cross-platform.yml

To run the Jupyter Notebook, simply open a command prompt in the forked repository, type Jupyter Notebook and then click gbif_analysis.ipynb.

To run the python code with the sample data, open a command prompt (Anaconda prompt) in the forked repository and then run

python gbif_analysis.py

It is using the configurations given in config.json. Alternatively, uncomment line 181 to provide your own configuration file via the command prompt. The following parameters can be provided in the configuration file:

• "species_list": List of species for which to conduct the analysis. Note that the first species is considered the target species for which to investigate co-occurence with the other species.
• "country_code": Country code to limit the analysis to a specific country only, e.g. 'DE' for Germany
• "output_directory": Absolute or relative output directory to save output CSV-files and PNG-files to.