Developing methodology for “smart” search for political process markers in social media

Abstract

The paper describes the methodological framework for political process studies based on social media markers. The empirical basis of the paper is a number of authors’ studies carried out using Predictor Mining, an interdisciplinary approach combining Data Mining and Social Computing. Predictor Mining is a computer-assisted technique which allows extracting data from unstructured chaotic large data flow for further analysis and prediction of political protests and their dynamics. A special attention is paid to the core principles of predictive mining related to the creation of protest markers of online user behavior in Russia and abroad. The article describes how Data Mining and Social Computing techniques help to evaluate the risks and possibilities to resist the technologies of mobilization of protest settings in the modern Russian and post-Soviet countries. The authors present the results of a comparative analysis of the specifics of mass political protest in Russia compared with Ukrainian and Moldavian cases. The following measurements related to the cases were made by the authors: analysis of active nihilistic and passive nihilistic data flows in the Ukrainian social media discourse (2014, computer-assisted classification of data flow); analysis of protest-related tag cloud in the Russian social media (2011—2017, discourse analysis); analysis of the use of social media as a tool for mobilizing political protest in Russia (2011—2017; cybermetrics); defining the structure of user roles and the density of social ties in the Russian and Ukrainian VKontakte protest communities (2017, building social graphs); analysis of social and political protest of collective actors in Russia (Manezh square, Kondopoga, Pugachev, Rostovon- Don, Biryulevo, Mineralnye Vody) (2006—2014, triangulation involving event analysis, discourse analysis and cybermetrics). Using a combination of different approaches to data acquisition and analysis such as survey techniques, content analysis, online and offline, qualitative and quantitative, traditional and Big Data techniques which are in fact a methodological experiment, the authors draw certain conclusions and propose recommendations related to the search for political protest markers in the web. As to the prospects of Predictor Mining, the authors consider that there is a need to elaborate a method that could encompass preliminary assessment, scenario planning and forecasting of regional and national social and political processes using the online traces.