For our project, we did a network analysis of a web forum at Firehouse.com, and then put this data into a Node XL graph. Our first task was to gather background information for each user that participated in our chosen thread, which is called "Why are cops on a power trip?". The data we collected for each user was length of membership to the site, and the avg. number of posts per day. We represent this data in our graph by having the size of each node relative to avg. posts per day (with larger nodes signifying more posts), and the color of each node relative to length of membership (with darker colors signifying longer membership).
After obtaining background profile information, we went through each post in the thread and recorded which user posted it, which user the post was directed at, and whether the subject of the post was supportive, negative, or indifferent towards the person it was directed towards. To show these variables in our graph we set our edges between nodes to be colored blue if the post was supportive, red if the post was negative, and purple if the post was indifferent.
An interesting observation from our graph is that if two users post negative remarks towards the same user, then posts between the two users will likely be supportive of one another.
*Note that CaptainGonzo and CrnkB8 both have red lines going towards Scarecrow, and a blue line inbetween them. This relationship triangle happens again between FireMedic, BULL321, and Bondga.
*Note that CaptainGonzo and CrnkB8 both have red lines going towards Scarecrow, and a blue line inbetween them. This relationship triangle happens again between FireMedic, BULL321, and Bondga.
Future work on this project could include analyzing more threads to follow user to user relationships and see if people agreeing on one topic are likely to agree on another.
Another interesting finding was ScareCrow was a fairly new member. He has the biggest post average in the post we looked at. And he also has the most negative comments directed at him. It would be interesting to follow his behavior and see if he creates same reaction out of members in different threads; negative comments directed towards him!
ReplyDeleteI would also be interesting to maybe look at other threads and find out a ratio or percentage of how many comments are negative,supportive or indifferent in a thread. This could perhaps tell us if for example if people are being negative based on the topic or is it just because they are trouble makers.
ReplyDelete