This repostory contains the prototype for a graphic user interface for the causal time series analysis tool tigramite.
This prototype is a module for JupyterNotebooks, so make sure to have a running version. Also, before using it, make sure to install tigramite on your device.
Tigramite depends on Python v3.7, to ensure compatibility, use also Python 3.7.
Also, it is necessary to have mpi.h.
For ubuntu systems, sudo apt install libopenmpi-dev mpi
.
In the past, using a conda environment for the prototype did not open the module. If your window looks like this after running the cell, install the package in your base python. Here, only numpy arrays with the ending .npy can be loaded in.
Follow these steps:
- clone this repository on your computer (e.g. <root>).
- in <root>
pip install .
After installation, open a new Jupyter Notebook or open the existing in the project. The following code will import and open the prototype.
from PrototypeWidgets import ProjectWindow
ProjectWindow().show()
It should look like this: With the lower left corner showing the algorith output and the lower right corner the plot output. To use it, just put your .npy data in \path and load it in. All defaults are set and can be changed through the accordion view.
Go to the directory examples, and open ExampleNotebook.ipynb. The .npy files in the same directory can be used to get to know the workflow.
Copyright (c) 2020-2021 Sarah Bachinger
See the LICENSE.txt file for more information. Also check the tigramite repository for their license.
- J. Runge, P. Nowack, M. Kretschmer, S. Flaxman, D. Sejdinovic, Detecting and quantifying causal associations in large nonlinear time series datasets. Sci. Adv. 5, eaau4996 (2019). https://advances.sciencemag.org/content/5/11/eaau4996
- J. Runge (2020): Discovering contemporaneous and lagged causal relations in autocorrelated nonlinear time series datasets. Proceedings of the 36th Conference on Uncertainty in Artificial Intelligence, UAI 2020,Toronto, Canada, 2019, AUAI Press, 2020. http://auai.org/uai2020/proceedings/579_main_paper.pdf
- J. Runge (2018): Causal Network Reconstruction from Time Series: From Theoretical Assumptions to Practical Estimation. Chaos: An Interdisciplinary Journal of Nonlinear Science 28 (7): 075310. https://aip.scitation.org/doi/10.1063/1.5025050 Nature Communications Perspective paper: https://www.nature.com/articles/s41467-019-10105-3
- J. Runge et al. (2015): Identifying causal gateways and mediators in complex spatio-temporal systems. Nature Communications, 6, 8502. http://doi.org/10.1038/ncomms9502
- J. Runge (2015): Quantifying information transfer and mediation along causal pathways in complex systems. Phys. Rev. E, 92(6), 62829. http://doi.org/10.1103/PhysRevE.92.062829
- J. Runge (2018): Conditional Independence Testing Based on a Nearest-Neighbor Estimator of Conditional Mutual Information. In Proceedings of the 21st International Conference on Artificial Intelligence and Statistics. http://proceedings.mlr.press/v84/runge18a.html