This repository holds a collection of python notebooks that can be used to replicate the findings of the paper Detecting climate teleconnections with Granger causality.
Filipi N Silva, , Didier A. Vega-Oliveros, Xiaoran Yan, Alessandro Flammini, Filippo Menczer, Filippo Radicchi, Ben Kravitz, and Santo Fortunato. "Detecting climate teleconnections with Granger causality." arXiv:2012.03848 [physics.ao-ph] (2020). http://arxiv.org/abs/2012.03848
- Filipi N. Silva
- Didier A. Vega-Oliveros
- Xiaoran Yan
- Alessandro Flammini
- Filippo Menczer
- Filippo Radicchi
- Ben Kravitz
- Santo Fortunato
- First download the preprocessed data from zenodo and save it in
DataFolder. - Use the
CalculateMetricsnotebook to calculate lagged correlation and Granger Causality. - Use the
ValidateMetricsnotebook to evaluate the metrics according to the El Ninõ and La Niña ground truths. - Use the
DrawMapsnotebook to construct a figure with the data. - You can load your own data by loading and converting it in the
ParseDatanotebook.
The Data folder contains example output data used in the article for Granger Causality (Causality_d7_l<maxLag>_v3.csv) and Lagged Correlation (Causality_d30_l<maxLag>_v3.csv).
The preprocessed data is avaiable on zenodo: https://dx.doi.org/10.5281/zenodo.4270623
Conda environment is recommended to run the current code. We provide an environment in environment.yml.
First, install conda or miniconda from https://docs.conda.io/en/latest/miniconda.html then run
conda env create -f environment.yml- Make it modular and release it as a pyPI package.
If you use the processed data or any of the scripts in a publication, please cite the respective works listed here.
Filipi N Silva, , Didier A. Vega-Oliveros, Xiaoran Yan, Alessandro Flammini, Filippo Menczer, Filippo Radicchi, Ben Kravitz, and Santo Fortunato. "Detecting climate teleconnections with Granger causality." arXiv:2012.03848 [physics.ao-ph] (2020). http://arxiv.org/abs/2012.03848