Big Data Applications in Social Media Research: The Experience of Unsuccessful Data Analysis of Ukrainian Facebook

Abstract

This article showcases a detailed description of the first stage of research on the discourse of Ukrainian opinion leaders on Facebook conducted by a team of researchers representing MGIMO University, Lomonosov Moscow State University, and Institute of Economy at the Russian Academy of Sciences. Convinced that it is Facebook that serves as the primary means of communication of politicians with the population in Ukraine, the team built a data base consisting of posts written over a 10-month period by 176 profiles belonging to the representatives of Ukrainian elites, and applied machine data analysis. The research question was the following: What strategies on the conflict in Donbas are verbalized by the Ukrainian elites? The author faced three challenges and limitations of machine data processing and analysis:unsuccessful operationalization of terms; functional limitations of the Semantic Archive Platform, of which the author turned out to have unreasonably high expectations; lack of understanding of peculiarities of Big Data analysis. Nevertheless, it was the failure of this pilot research that helped raise crucial questions for further research, primarily on the criteria for shaping a data base and on formulating of the research questions for software. This experience turned to be essential for second and third stages of the research project that were completed a year and half after the project was launched. Hence the necessity to make public all the considerations on this research.