I hope we are still allowed to make mistakes, because this post is about mine in Gephi. I made a network visualization of some tweets related to the Boston Marathon bombings, but I couldn't seem to make my analysis meaningful. I followed Grandjean's helpful article and waited a very long time to make the picture below. But when I tried to make source nodes, Gephi couldn't find paths between my nodes. I realize that I am probably doing this wrong, and there is maybe an easy solution, or maybe my data set isn't a good one in some sense. Below are my mistakes; below them further are what I wanted to achieve (and will achieve in time!):

I wanted to find out if, sticking with the same question as last week, I could use social network analysis to trace how information was spread via followers on Twitter in order to ask the question "can social network analysis make Twitter more reliable in cases of disaster awareness?". For example, I wanted to find out: can anyone spread information that can become viral related to disaster awareness and prediction? I was thinking, again, of the Sunil Tripathi, the Brown university student falsely identified as a suspect via Reddit and Twitter. If we could know how information was spread on Twitter in this case--how we could reimagine Twitter top trends lists, for example--could we make Twitter more reliable in cases where the blurring between facts, rumors, and fictions are often destructively intertwined? I wanted to make a social network analysis of Twitter relationships like Eric's one below:

Eric used Gephi's modularity routines to color-code different communities. He identifies them as: data science (green), general tech (yellow), general news (blue), astronomy and webcomics (teal), and pro cycling (maroon). The disconnected components are food and Cal football. How can we map "failed" Twitter communication that has resulted in misidentifying suspects of crimes? How could this help Twitter reimagine its current structure?