Aim is to analyze the textual data, comprising of the tweets of the politicians in order to gain insights and patterns from such social media driven conversations for the typical conflicts that very quickly escalate into online hatred.
To observe and hence quantify the effects of polarization and hatred from the tweets of some of the politicians of The United States of America.
- Forming better KPIs for dealing with Polarization and groupthink when it comes to Big Data.
- How does Big Data let you generalize better on the social network? Or is doing a fragment-based analysis helpful?
- Can we identify and hence quantify the “tipping point” in a series of conversations (maybe from politics - depends on data) that we could use to prevent online disagreement or hatred. - Advanced goals include suggesting a mid-way “negotiation” in order to alleviate the disagreement.
- Our approach for the extraction of the conversations for the study.
- Unsupervisedly cluster the group of conversations - Maybe a dendogram is a good choice
- Are the well-informed the source of the conflict or the ill-informed ones? Can such insouciance lead to one being a victim of Groupthink? - How to determine this via the short conversations/sentences?
- Does the temperament, interest, discomfort of the conservation imply any correlation with such biases? Being anonymous leads to more such presumptuous/controversial opinions?
- Being highly controversial leads to an increase in the Word Of Mouth. Does this have an effect on people choosing sides?
- Will learning how the political graph(a social network among the actors of the conversation) of such conversations can be helpful?
- Does time play any role? Old comments or newer comments?
- Can this be modelled as a Makov Chain instead of NNs