2
$\begingroup$

I recently saw the graph of correlated technologies from the SO Developer Survey, which was pretty; reproduced below:

SO Developer Survey correlated technologies

I have a weighted graph that represents the amount of communication between pairs of people. I'd like to create a similar visualization as above. I want to be able to tease out "islands" of people with high influence among each other, but little influence elsewhere.

To make this concrete, let's say I have four animals: dog, wolf, cat, lion. The canines talk a lot amongst themselves; the felines do too, but members of one group only rarely talk to members of the other group. So my influence graph (which is symmetric, hence the upper triangular matrix) looks like:

+------+-----+------+-----+------+
|      | Dog | Wolf | Cat | Lion |
+------+-----+------+-----+------+
| Dog  | ∞   | 100  | 1   | 1    |
| Wolf | -   | ∞    | 3   | 2    |
| Cat  | -   | -    | ∞   | 130  |
| Lion | -   | -    | -   | ∞    |
+------+-----+------+-----+------+

I know how to create graphs whose edges are visually represented by the weight between two nodes, but I am not sure how to account or calculate islands of influence. Are there general visualization techniques to apply to a data set like this, and also for figuring out islands of influence?

Thanks in advance!

$\endgroup$
  • 2
    $\begingroup$ Welcome to the site! Read about community detection. $\endgroup$ – Emre Sep 15 '17 at 21:28
  • $\begingroup$ Yes, the answer to your question was given in the comment above. I totally agree with Emre. $\endgroup$ – Pablo Suau Sep 22 '17 at 12:56
  • $\begingroup$ Great, it has definitely helped. I can't accept a comment so @Emre can you write that as an answer? $\endgroup$ – ash Sep 25 '17 at 11:22
2
$\begingroup$

As you've explained it, you want to detect communities, which is a common problem in social network analysis. If your network had been directed and temporal, I would have suggested you look into influence analysis; e.g., as in Detecting Communities of Authority and Analyzing Their Influence in Dynamic Social Networks.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.