Insights from social networks: a big data analytics approach.

Abstract

One of the biggest domains of insights of Big Data are online social networks(OSN), whose paths for knowledge are currently under exploration. The unfolding of every event, breaking new or trend flows in real time inside OSN triggering a surge of opinionated networked content. Such unprecedented scale of human communication and public behaviour data brings new opportunities to understand how society works. Tools are fundamental to help people perform data analysis tasks. However, meager progress is done; not only because science is still far from automatically processing human-centric data but contextual concerns such as privacy or noise restricts this endeavour as well. OSNs aren’t yet examined at the cleansing stage and until now data analysis has been studied separately. With this research we are trying to make the start. Investigating the social networks as a contextual source, their data value chain, the low quality data they contain and how could these be addressed, constitutes the first step. Secondly, the entire spectrum of social networking data analysis, namely (i) social network analysis, (ii) sentiment analysis, (iii) topic detection and (iv)collaborative recommendation is studied. In particular, our purpose is to exploit state-of the-art frameworks and techniques and correlate them in both cleansing and analysis services. Then, we develop a cross-platform tool for sentiment analysis through modern cloud-hosted machine learning services. Lastly, the idea is to capture both analysis limitations and future trends regarding data from OSN with a special interest on sentiment analysis and computational intelligence paradigm.