I chose 3,500 tweets as my data set from the days March 1st to February 24th. The reason I chose to do the maximum number of tweets is because of the amount of data I was able to collect in the past month. The reason I picked these days was because I wanted to analyze more recent data as I analyzed the days after the muslim ban was first implemented for my text analysis.
I chose to run number 1 which means that it would not differentiate between retweets and non retweets, including singletons. I wasn’t too sure what to expect from this type of extraction.
After opening the Les Miserables data in gephi it indicated that there are 77 nodes and 254 edges. I chose to do an undirected graph because I would suspect that a novel would contain a lot of mutual relationships such a friendship or dating and thought it would be interesting to visually see this relationships. Valijean’s ID is 11 and Fantine’s ID is 23. From this I was able to figure out that Valijean talks to 4 different characters while Fantine talks to nine.
When I increased the repulsion strength from 200 to 10,000 I was able to see the connections between each character better due to the nodes wanting to spread out.
The average pathway length was 2.641 and the network diameter was 6.
After changing the betweenness centrality to a minimum size of 10 and a maximum size of 200, several of the nodes on my graph became much bigger. I believe this happened because the ones that increased the greatest are the characters who have the most connections.
Analysis of my data
I chose a 10,000 repulsion strength because I felt this would enable my data to spread out enough in order to see the connections between my tweets. My average path length was 7.174 which means there is about a 7 degrees of separation in my data set between people and the network diameter is 20 which means that people are usually 20 connections apart.
I thought it was interesting that the biggest node on my graph represented the realdonaldtrump. I noticed after referring back to my tweets that most people were tweeting at him in a negative connotation. I also think it is interesting that the realdonaldtrump node was bigger than the POTUS node because POTUS is the official Twitter of the president. The next node that had great significance represented the handle funder. I wasn’t too sure who this could be so I looked up the handle. It turns out his name is Scoot Sworkin and he is the co founder of the democratic coalition against trump. His tweet that was getting the most attention was “RT if you think trumps #muslimban was a disaster #trumprussia.” In comparison to my text analysis I thought this was interesting as it was talking about the after math of the muslimban rather than people’s opinions when it was first implemented. Another node that was fairly big represented the handle RVAwonk. I was curious to see who this was and after looking it up it turned out to be a women named Caroline. Her bio states that she is a behavioral scientist and a long term trump critic. Her tweet that was getting the most attention stated “Read this -> Stephen Miller just admitted that Trump’s #MuslimBan isn’t driven by concerns about national security.” Following her tweet she added a link to a Washington post article. Another thing I found note worthy was the amount of connections between realdonaldtrump and nancypelosi. Her most popular tweet during this time said “@realDonaldTrump’s #MuslimBan dishonors our values, violates our Constitution and undermines America’s fight against terror.” This shows that even politicians were still tweeting about Trump’s policy even a month after it was issued.
When looking at my graph I had a significant number of tweets that weren’t connected. These tweets can be shown by all of the gray nodes floating around on my graph. I think this is due the fact that I have such big data in relation to my hashtag. I would be interested to see what all 84,000 of my tweets would look like if I graphed them as I think I would have a vast amount of connections. However, at the same time I would need to weed out all of the unrelated tweets in order to tell my story better. After looking at my graph I was also able to discover that more people are vocal when it comes to being against the muslim ban then for it. Overall I enjoyed this lab as I was able to see my data come alive.
As usual, we had a lot of similarities in the analysis of our tweets. I’m not surprised to see that realdonaldtrump was the most prominent tweeter in your visualization, however I did find it interesting that there was predominantly negative feedback in the tweets. This is likely because if someone were a supporter of the muslim ban, it’s likely that they would not still be discussing it all that frequently. It would be interesting to see how the tweets regarding the muslim ban changed over the span of time from the chemical attack in Syria, through the U.S. bombing of the military base last night. I think your choice of repulsion rate was a very good one, as it allows you to really see who is the most prominent in the discussions of #muslimban. Overall, great work with some cool results.