ewstools: A Python package for early warning signals of bifurcations in time series data

Abstract

ewstools is an accessible Python package for computing, analysing, and visualising EWS in time series data. The package provides: • An intuitive, object-oriented framework for working with EWS in a given time series • A suite of temporal EWS and associated methods (Dakos et al., 2012) • A suite of spectral EWS (Bury et al., 2020) • Methods to use deep learning classifiers for EWS (Bury et al., 2021) • Integrated plotting and evaluation functions to quickly check performance of EWS • Built-in theoretical models to test EWS • Interactive tutorials in the form of Jupyter notebooks