This python implementation can be used to reproduce the figures from my essay submitted for MA4M4, titled "Using asymptotic surprise to discover communities in a sea surface temperature network". The file main.py is the entry point, while all 'content' can be found in the ma4m4 package. The file environment.yml can be used to reproduce the conda the virtual environment:

conda env create -n ma4m4-jb --file environment.yaml

Note that ma4m4 has not been properly packaged as a python package, so is not installed. Instead, main.py must be run from the project root directory, so that ma4m4 appears on the python path.

python main.py

Note that you must download the data before you can run the pipeline (see below).

The signal processing to generate the anomaly series and correlations is done using numpy and scipy. Graphs are represented using networkx and cdlib is used to perform all community detection. Plots are generated using matplotlib and cartopy.

Data can be downloaded in NetCDF format from https://www.metoffice.gov.uk/hadobs/hadisst/data/download.html . The file should be saved as HadISST_sst.nc in data/01_raw/.