Social Network Analysis

(Jon Richter) #1

Today I have received an enthusiastic message from @kevflanagan who is currently investigating Social Network Analysis with Gephi. I am allowing myself to replicate this message here and open it to a public for broader discussion.

Have you ever played around with Social Network Analysis?

I’m learning how to do this for my PhD. Playing around with Gephi and harvesting facebook and twitter data. More or less inspired after reading the attached article Online network organization of Barcelona.pdf (3.3 MB).


One of the things that I was thinking about after the inter-mapping was the difference between formal and informal mapping processes. In a sense much of what we were talking about falls into the mapping of formal relationships, voluntarily contributed. What SNA does is map informal relationships. What could we learn from the differences? The impression I get from reading the attached article is that the researchers have been involved in using SNA as part of a reflexive process to answer questions such as ‘How are social movements connected with political groups?’, “How can those connections be strengthened without compromising the integrity of the movements?”. I have a graph with over 50,000 facebook page nodes. My computer is slow as shit. But the ‘community detection’ algorithm which has no knowledge of geography or language was able by math alone to identify 15 groups which I can see roughly correspond to different countries and language groups. I’ve become a bit addicted to this over the past month. Would love to hear your thoughts on this stuff particularly as it relates to mapping and supporting relationship between commons groups.

At first I have only been thinking of @Amaury_Van_Espen’s and @daniel’s experiments in

but recently also informed few Degrowth researches about
What stroke me in reading the PDF was Fig. 11, which resembled so much the loop-diagrams from

We also know that the Edgeryders actively use Edgesense in their work to follow their communities (demo I, demo II).

Please note such an online map may draw a false image of them, as it only, and only contains online interaction. This slightly touches the fallacies with mapping we may find when following the research questions that @tuomo recently brought up in

Did you two converse yet, too?
Unfortunately I personally don’t have any experience in exploring these kinds of datasets, but maybe a wider crowd knows further answers?

(Tuomo Alhojärvi) #2

Hi there,

Oh yes, I agree. At least to the extent I understand what’s going on here. I think there’s definitely a burden or two that come with the technologies we use, their constant re-envisioning notwithstanding, but there are different degrees of awareness about those baggages, and capacities to act on them, which makes all the difference.

Nope, not yet. I think I still have an email pending somewhere in @kevflanagan’s inbox… :wink: To be honest, I know very little about social network analyses, but very much interested to learn more!

(Kevin Flanagan) #3

Ok I found your email @tuomo :slight_smile:

Yes. Let’s arrange a call.

I propose a small weekend project. @almereyda or @tuomo is there a particular community you know that is well documented on Transformap? What I propose is to do a small social network analysis project basically harvesting facebook page data through netvizz and visualising it with gephi, creating a graphic and then comparing with the Transformap and discussing the results in a blog post?

It would be a weekend project. I can’t really give it much more time than that these days unfortunately.