Predicting social unrest via textual analytics
The aim of this project is to develop a method to predict the likelihood of social unrest occurring in a given location by using alternative and economic data.
Project background and aims
Social unrest impacts societies all over the globe. Causes of unrest are many and varied, but are often triggered by either machinations in the political class, or basic Economic factors. Unfortunately, in many authoritarian or autocratic regimes, social unrest leads to violence, destruction of property, and sadly physical harm and death.
The aim of this project is to develop a method to predict the likelihood of social unrest occurring in a given location by using alternative and economic data. Using alternative data from the GDELT textual analysis project, together with economic series from the OECD and other sources, we use AI methods to predict the propensity of social unrest one to four weeks in advance.
It is anticipated that such a method will provide the international community with much needed forewarning of likely flash-points around the globe, such that diplomatic, observer, and other actions can be taken to diffuse the crisis, protect citizens, and document any resulting harms.