2 years ago
TechAtHeart
in English · 15,480 Views
likes 6clips 4comments 4
A Network Analysis of #Gamergate
Chris von Csefalvay is an expert technologist specialized in big data that posted an article yesterday that has been getting much attention. He did a network analysis on a sample of 30,000 tweets (about 5K per day) from December 1st to December 6th 2014 that used the #Gamergate hashtag. It is an interesting read. He did clustering methods in order to analyze the data and then did network visualization to produce the graphs. The first picture I attached is focused on the nodes with a degree equal to or greater than the average. There is a visible center even though the #gamergate network is "leaderless". Almost 35% of the nodes and 3/4 of the node edges are part of the whole "giant component", explains von Csefalvay in is his article. Here is part of his conclusion after all his findings: "Instead of a concentric, hub-and-spoke pattern, in which a few personalities emerge as leading the field and engage/are engaged by others, in the case of #Gamergate, it's small community clusters that lead the field. The participants are strongly interconnected, but thanks to the hashtag's ubiquity, most users seem to be open to contact and interaction with most other users, leading to a diffuse and weakly linked structure. For a political/consumer pressure group, this is definitely a clear advantage." Of course, all of these findings are not able to prove or disprove if Gamergate is a hate group or not. It's still a pretty interesting analysis. Make sure to read his full post to get a full grasp of his analysis. What is everyone's thought on this? (via http://chrisvoncsefalvay.com/2014/12/07/Gamergate.html)
TechAtHeart clipped in 1 collections
4 comments
This is so fascinating! Thank you for sharing it, @TechAtHeart. I have so many feelings about Gamergate, but leave it to amazing data analysts to put it into a way that I can step back and see from a different perspective!
Oh, and do you mind adding this to Video Games as well? Thanks so much! I think people there will be really interested.
@sanityscout, same here. It's interesting to have another perspective and see relationships and connections that we can't see without these type of experiments. There's so much to learn, it can be so overwhelming lol. I'll make sure to add it, thanks for letting me know!
Wow, this is really impressive, thanks for sharing @TechAtHeart. Some of it goes over my head...but I'm so glad he was able to explain how he parsed his data with some of those code snippets.