First of all, Google+ must be applauded for attacking the delicate problem of controlling who sees what of the information to put online. The concept of circles isn’t significantly different from Facebook’s friend groups – but Google+ makes this aspect central both when creating new contacts and when posting new content. This is more than an interface tweak – defaults matter (*). As a result, the semantics of Google+ “friendships” in my opinion already differ from those on Facebook, seeing how people add me to their circles who would never request a Facebook friendship (e.g., unknown friends of friends). However, this new-found privacy may lull people into a false sense of security: Excluding someone in your circles from something you share does not mean they won’t ever see it. They just won’t see it from you.

This may be obvious. It wasn’t for me, so I thought I’ll share a little experiment. Consider you have two friends, A – your friend, and B – your business contact. You share an update which is only intended for your friends, so you exclude the circle including B. A finds this so hilarious he shares it with all of his contacts. These contacts happen to include – guess what – your business aquaintance B. Will B see your update? Yes, she will. You may think “but I told Google+ I don’t want her to see this”. No, you didn’t! You just stated that she won’t be in the set of people who will receive that information from you (**). Any attachment you put on that post will be generously forwarded to B as well. To add further awkwardness, B will see that A was included in the circles of the original post, but she wasn’t, as she’s getting a “A shared [your name] post with you” message. Ouch! In Google+’s defense, you have to say that a) re-sharers get a pop-up telling them that the original post was shared in a limited scope and b) you can disable re-sharing. However, the re-sharer may not care, and the re-share disabling is an option available in a drop-down menu which appears after you post (i.e. much less prominent than the circle selection, and we know what happens to non-default functions).

To conclude: Circles may be misunderstood by some users (e.g., me – I’d love some comments from readers telling me how far off I am here) as serving a purpose which, in fact, can only achieved in a brute-force way via “disable re-sharing”. Solutions would include users being aware of this, Google+ deemphasizing the privacy power of circles, and/or making the “disable re-share” button more salient (or change the semantics to “people I haven’t shared this with are excluded from future re-shares”). Of course: Privacy is hard, and particularly in a social network, giving up some of your privacy is part of why it works. All the more, it’s important for existing privacy features to be clear, otherwise they may cause problems you never would’ve had, had you simply not relied on them at all and just applied the basic heuristic for putting stuff on the internet: Don’t post anything you wouldn’t like to see on the cover of tomorrow’s newspaper!

This post is based on a discussion with Oli T. and stars Crisi K. in the role of B.

(*) Which is why Circle Hack, a Facebook app copying Google+’s interface for assigning friends into groups, doesn’t really close the gap between the two.

(**) It does make sense, not stating X is something different than stating that not X, just be clear that removing a circle does the former.

Update: At least one other person finds this behaviour noteworthy as well (in German)

In my PhD thesis, I try to make sense of the large bodies of data that are accumulated by people saving resources online, tagging them with whatever words they choose in order to find them later on. Each time a user u saves a document d, using the tags t1, t2, and t3, three triples (d,u,t1), (d,u,t2), (d,u,t3) are created. In this way, the users of large social bookmarking sites like Delicious have created datasets of several hundreds of millions of triples. Over the last years, a whole body of literature has been created that’s concerned with making sense of this data (there’s an intuition that something valuable is in there; after all, each of these millions of triples means that somebody has thought *something*!). Here, I want to show a few visualizations of various social bookmarking datasets aimed to provide a quick idea of the complexity and the approximate content of these networks.

Here’s what I’ve done:

• Take the 1000 most popular documents from each dataset
• Compute the similarity between them in terms of associated tags
• Draw connections between the 3000 closest pairs of documents
• Draw connections between each document and its most frequently associated tag
• Scale the tags by frequency
• Remove everything that’s not connected to the largest component of the graph
• Run the resulting network through GraphViz, a fantastic graph visualization software

For the technically inclined, the similarity between documents is computed as a so-called cosine similarity: Each document is represented by a large vector containing, for each tag, the number of users that have used that tag to describe the document. The cosine similarity between those vectors is then a number between 0 (no similarity) and 1 (complete match). If you imagine two vectors (0,1) and (1,0), the angle between them is 90 degrees and the cosine of that angle is 0. This can be scaled up to many dimensions and is frequently used in information retrieval as a basic similarity measure between measure. Also, the choice of the numbers 1000 and 3000 is somewhat arbitrary. They just happen to create what I find the most expressive visualizations over all datasets. The actual visualization, as noted, is outsourced to GraphViz; for a quick intuition about what it does, just imagine that each connection is a spring between two nodes – the visualization is a stable state of the resulting system where the different spring forces are in balance.

Visualize.us, a social bookmarking site for images.

One month of Delicious (12/2007) [2], a social bookmarking service for URLs

CiteULike [3], a social bookmarking service for scientific articles

Bibsonomy [4], a social bookmarking service for scientific articles and URLS

Bibsonomy, without the spam removed

Through the way they are created, these visualizations are heavily biased towards the most popular documents, so a lot of content is lost – the original point of these visualizations was just to demonstrate the brutal way in which spammers destroy inherent patterns in the data (even though I think there’s more things to find). For work more aimed at a complete overview of a particular dataset, please check this nice visualization[5].

Higher resolution or vectorized versions available on request; if you would like to cite this work in an academic context, please refer to [1]

[1] Nicolas Neubauer and Klaus Obermayer: Hyperincident Connected Components in Tagging Networks. In Proceedings of the 20th ACM Conference on Hypertext and Hypermedia [pdf]
[2] Robert Wetzker, Carsten Zimmermann, Christian Bauckhage: Analyzing Social Bookmarking Systems: A del.icio.us Cookbook. In Proceedings of the ECAI 2008 Mining Social Data Workshop (2008), pp. 26-30.
[3] http://www.citeulike.org/faq/data.adp
[4] Knowledge & Data Engineering Group, University of Kassel Benchmark folksonomy data from bibsonomy, version of june 30th, 2008.
[5] Nianli Ma, Russell J. Duhon, Elisha F. Hardy, Katy Börner (2009) Bibsonomy Anatomy, Sunbelt Viszards Map. Online at http://cns.iu.edu/research/09-Bibsomony.jpg

Graphical Recording of my talk at WIN'10. (by AlphaChimps)

A few days ago, I returned from “Workshop in Information in Networks”, where I had presented our paper Community Detection in Tagging-Induced Hypergraphs. It was a two-day workshop with an informal feel to it, celebrating interdisciplinarity and featuring research from physicists and historians, computer and political scientists all while being hosted at NYU’s Stern School of Business.

Short snapshot of what comes to my mind after almost two weeks now after browsing the program again (citations without guarantees):

• Duncan Watts on cascading in Twitter: “Almost nothing cascades”
• Nathan Eagle on real-world relevance of research: “The really hard questions come from people who want to know where to build a fountain to minimize diseases” (see also AI for Development – http://ai-d.org/)
• Stanley Wasserman polarizing on network researchers neglecting the roots of their field and hypothesis-driven (statistics) vs data-driven (data mining) research
• Jeff Nickerson employing Mechanical Turk for crowdsourced brainstorming
• Isabelle Stanton demonstrating the construction of random graphs with a given joint degree distribution
• Michael Kearns on coordination problems in networks, showing through elegant experiments how, e.g., “a well-connected, small elite can enforce their will against an unconnected majority”
• Neha Gondal showing how a network dataset of loans in renaissance Florance could be used to test various theories of the social dynamics of lending in that time
• Peter Boothe showing that 18 autonomous systems (ie ISPs etc) could together kill around 40% (or at least some ridiculously high number around there) of all internet traffic by turning off their systems
• Sinan Aral showing that susceptibility to peer pressure (as measured by following a Facebook recommendation) increases as the relationship status changes from single to married (most susceptible, though, are “it’s complicated”-customers)

One nice feature of the event was the facilitator/graphical recorder from http://www.alphachimp.com who made funny but amazingly effective sketches of the talks for documentation and later discussion. Above you find the recording of my talk, linking to the full set of recordings from the event at Flickr.

Btw you will find I put some advertising on the page. I wasn’t originally planning on this and also don’t expect to become rich by this in the near future, but the slick option editor of WordPress and this beautiful theme called http://www.khairul-syahir.com/wordpress-dev/graphene-theme made it so simple I just had to try it out. Ah, well…. (same goes for all the social media bling ;)