Join me on Jabber using matthias_studer [at] jabber [dot] org or Contact me
If you want to know wich node you are
You want me to do a special network analysis (or for one special mailing list)
If you want to make comments
Some basic of interpretation
First of all, Social network graph are just (and not more than) a presumption of a social structure. There may be wrong, since I don't know the social structure from inside. Therefore, this networks are only probabilistic!
Second, there are not a friendship network, rather a collaboration network (there may be some big differences). This is due to the available information.
How relation and network are built
In the graph showned only what we can consider as the center of the network is showed (otherwise graph would be too big).
We have defined the relation in the following way :
A message is considered sent to all participant of a given thread. The underlying reason is: if you participate in the discussion, there are more chance you consider and collaborate with other participant (therefore it's probabilistic)
The relation is then valued in the following way (it's a quite usual way to do it): if Mij is the number of message sent by individual i to individual j then the relation between i and j equal sqrt(Mij*Mji).
This give the following properties: A relation of 6 can be either each individual sending 6 message (sqrt(6*6)=6) or one individual sending 4 and the otherone 9 (sqrt(9*4)=6). Thus we considere that inequal relation are less important. We want to consider and include strong asymetrical relations since there may be important in the social network. Thus a node that appear on the network doesn't mean that he is a great contributor, but that he is important to consider the social structure.
The graph are built in the following way. We consider only individual participating to at least 2 threads (others are disregarded). We consider only relations greather than or equal 6 (it's a kind of mean for one message from each individual per month) only people that have at least one relation are shown (again otherwise it would be much bigger).
We then use a log transformation (that lower the difference between high relations)
According to my data, the relation of "6" condition is a restrictive one.
Node size are comparable inside a given graph (not accross) and size are only meaningfull in relation to size of other nodes.
It is not because a node is on the center that is a more central one (actually on mailing lists central nodes have a different meaning than on usual social relations, because everyone can contact everyone).
You should always check the web of relations and the strong of relationships (darker lines). This is more informative. The centrality indice computed is based on number of relations.
After saying that, just another thing. We do not think that a greater number of message, a greater number of commit or whatever is better. In our sens it reflect (in a probabilistic way) different social positions. That's why we have a questionnaire in relation to this.
Warning: include(comment.php) [function.include]: failed to open stream: No such file or directory in /home/www/1224d42a224aba6ed59ff8c9552db9c3/web/kdesurvey/sna/index.php on line 62
Warning: include(comment.php) [function.include]: failed to open stream: No such file or directory in /home/www/1224d42a224aba6ed59ff8c9552db9c3/web/kdesurvey/sna/index.php on line 62
Warning: include() [function.include]: Failed opening 'comment.php' for inclusion (include_path='.:/opt/php/lib/php') in /home/www/1224d42a224aba6ed59ff8c9552db9c3/web/kdesurvey/sna/index.php on line 62