Early Warning Signals in Open Source Intelligence: Two Use Cases of the 2019 Iraqi and 2020 Indian Farmers' Protests

Abstract

Early warning signals methods have been introduced in the field of ecological sciences and widely used in other domains. However, while these methods have proven effective for deterministic dynamics governed by differential equations or smooth maps – both on synthetic and real data – their application in the social sciences is more complex. A series of protests started in Iraq on 1 October 2019 and farmers' protests in India in September 2020. We investigate in this work how these waves could have been anticipated using early warning signals for the time series of daily occurrences of protests. We use to this end metric-based indicators (autocorrelation at-lag-1, standard deviation and skewness), analyse trends using Kendall rank correlation and use bootstrapping methods to implement a statistical test exhibiting a regime shift (tipping points) in the dynamics of protests. We moreover highlight the importance of the standard deviation as an indicator.