Cloud-based architectures for geo-located blogosphere dynamics detection

Abstract

Social networking data threads emerge rapidly and such crowd-driven big data streams are valuable for detecting trends and opinions. For such analytics, conventional data mining approaches are challenged by both high-dimensionality and scalability concerns. Here, we leverage on the Cloud4Trends framework for collecting and analyzing geo-located microblogging content, partitioned into clusters under cloud-based infrastructures. Different cloud architectures are proposed to offer flexible solutions for geo-located data analytics with emphasis on incremental trend analysis. The proposed architectures are largely based on a set of service modules which facilitate the deployment of the experimentation on cloud infrastructures. Several experimentation remarks are highlighted to showcase the requirements and testing capabilities of different cloud computing settings. Keywords: social networks and wisdom of the crowd, geo-located blogosphere dynamics, social geo-located data clustering, cloud service deployment.