Inferring topics within social networking big data, towards an alternative for Socio-political measurement

Abstract

This research sought to measure some socio-political indicators using millions of opinionated messages from social network sourced big data. Thus, and using an enhanced mixed method for sentiment analysis and a fusion model algorithm to infer topics from short text, this study attempted to demonstrate the value of computational approaches in measuring some phenomena in the real social world and quantifying public opinion fluctuations in response to certain socio-political issues. The validity of the experimental results was examined by comparing them with data obtained from representative surveys, thus providing a better understanding of the relationships between online and offline opinion dynamics. This contribution is intended to be multidisciplinary, both useful for policymakers and opinion analysts to explore public trends and to inquire into sociopolitical issues.