Trick is to use the Cytoscape desktop application to experiment with and discover how to fully use the cytoscape environment to produce the maps. So far we have found that many of the core features in the cytoscape mapping environment are available in the js library. Second trick is to optimally exploit the ossn network data. The maps we are generating based on our current algorithm is producing edges only where an actual connection exists AND produces two edges with directional arrows if the connection is mutual.

I had a first brief look into the basics and did that circle of Botox' friends. Really amazing. :)

... still a smart algorithm missing to avoid unnecessary crossings ...
as in fact it could look like this:

Our use case for network mapping is twofold. First, we want to be able to dynamically map the entire instance of the OSSN network for administrative purposes. Second, and more importantly, we plan to use individual user network mapping to provide a user with a visualization of his/her social network in order to further encourage the user to expand their network strategically. We are creating network wide and user specific maps now based on the user, relationships and object tables. Next step is to curate the strength of the relationship and use that data to modify the edge attributes. Just to ground the discussion ... a social network map looks like but in our use case is specific to OSSN users and groups.

Working? No. Interested: Yes.