TrackInfo: Finding Relevant Information from High Velocity Data of Social Network

Social networking sites are becoming popular platforms to share information for different types of media agencies including Newspapers, TV channels, FM radios, etc. Obviously, they hold public pages or accounts in social networks. Sometimes, such media agencies post in online social network (OSN) (Facebook, Twitter, etc.) and ask users' feedback through comments. They also declare some lucrative facilities for a user who give maximum comments to that post or contact directly to the users who give qualitative comments. In case of a live reality shows in TV channels, they ask votes from viewers for a fixed set of candidates within very limited time. As a result, in few minutes, more than 1,000 comments are found in such posts which is really difficult for streaming and analyzing to discover some information at the same time. In this paper, we discuss three new problems of such media agencies, analyze some scenarios and propose novel solution for each of the problem. Though such problems have not been tackled in the literature, we are first to deal with such problems. We have conducted an extensive set of experiments on the collected data from official facebook page of Roger Federer(1) and DhakaFM90.4(2) to evaluate the effectiveness of our proposed approaches.