This is a collection of Python scripts that I used while writing my Bachelor's thesis "Analysis of Spatial Patterns in STARS-Generated Data using Complex Networks and Kappa-Statistics" at the Potsdam-Institute for Climate Impact Research in 2016 under the supervision of Dr. Norbert Marwan and Prof. Jürgen Kurths.
The STARS model is developed at PIK, more details can be found here. My analysis of the STARS-generated hindcast dataset can be found in my thesis.
Feel free to contact me in case of any questions.
The scripts directory contains the scripts that can be used to handle STARS and BASIS data. The basic workflow can be seen in scripts/main.py. It consists of three steps:
- Structuring and preprocessing of the raw data
- Computation of Tsonis Climate Networks and Event Synchronization Networks and the associated network measures Degree, Strength and Directionality.
- Calculation of Cohen's Kappa between the network measure vectors.
Please remember to update the scripts/misc/paths.py to match your local directory structure.
The files scripts/misc/functions.py and scripts/misc/plotting.py provide some additional functionalities that might be useful for others.
The meta directory contains the meta data file I used while handling the STARS and BASIS data, it reflects my particular selection of meteorological stations in the dataset.
The pkg directory contains work provided by Aljoscha Rheinwalt that was used to calculate Event Synchronization Networks and compute the Strength and Directionality measures.
The project is written in Python 2.
In addition to the standard libraries like Numpy, Scipy, Matplotlib, Progressbar and Pandas, the package 'NetworkMeasures' by Aljoscha Rheinwalt has to be installed, see pkg/spat_corr.
The scripts expect the raw data in one single folder per dataset containing one file per station, e.g. a number of (id)simsz.dat files for STARS data and (id)basz.dat for BASIS data.