Prescriptive Analytics using the study of Network Science

Visualization: Mentions Network, Dataset: Twitter (#GamerGate November 2017)

Playgrounds: Python, Twitter, Gephi

The above ‘galaxy and stars’ fashioned visual is a deeply entangled directed network of Twitter users being targetted by all sorts of sentiments. On zooming in, we see the colors of edges representing the nature of a Tweet, and some hubs with a centrally located user (which I call a planet node) being targetted by what I call as satellite nodes.

0-4Also, the evolution shows loosely connected users which are either the newly introduced nodes (merging in) or dying nodes (drifting away) forming a ring around the network. Putting the centrality measures of Betweenness and Eigenvector values together comes along an interesting theory of identifying our star-players (promoters) and thus forecasting the evolution of a network by little steps on Predictive Analytics. All in all, an interesting study that can be used to both break or boost a network, thereby giving us methods towards both prevention and promotion of activities over the social web.

The first steps count towards the project contribution for the class of Network Science conducted by Dr. Mirsad Hadzikadic, now an continued ongoing research study with Dr. Wlodek Zadrozny’s assistance. Publications and detailed reports on the way. #DataScience #NetworkScience #Python #Twitter #Analytics

Leave a comment