Below is an experiment in using client side image mapping to creat an interactive map of the blogosphere. The idea is pretty straight forward - browse the graph with the mouse (note the URL associated with a node - i.e. a blog - in your browser) and click to visit the blog. This is taken directly from an idea from SOB.
Note that the map below was generated from data created for the WWE 2006 Workshop, and is from June 2005. Consequently, there may be links to blogs that don't exist any more - unlikely but possible. Note also that the map differs from others presented in the Mapping the Blogosphere Gallery in that I have used a higher threshold to define relationships between blogs (requiring 2 or more citations, not just 1). I did this to reduce the size of the graph for the purpose of tihs initial experiment.
To get started, simply mouse over some area and get oriented (the big orange node/blog is BoingBoing, for example). From there, drift around and click through. Depending on how things go, I will consider putting up a larger map of this and other communities.
Hi!
I'm looking your Map of the Blogosphere, and i want to invite you to our little map in our web [http://aseisgradosdeti.blogspot.com]. Our graphic is about phisical conections of the blogers. And our system of conections is diferent.
I enjoy with your map, and i want to give u congratulations for this work.
Bye!
Posted by: yuyu | July 04, 2006 at 03:49 AM
How are you positioning the nodes on these graphs? Are you using vertex elasticity, or is there some meaning to their position on the plane?
Posted by: Hans Gerwitz | July 10, 2006 at 10:28 AM
Sure this makes sense for some people but not for me. Cool pic though. ;o)
Posted by: Lautreamont | July 11, 2006 at 01:00 AM
Hans,
I'm using a variant of the force layout approach to graph layout. There certainly is meaning to the location of nodes in the image: proximity indicates a tendancy for mutual citation.
Posted by: Matthew Hurst | July 11, 2006 at 01:31 AM
Of course, I meant vertex repulsion/edge elasticity. You must have some nice resources to play with if you can afford force layout on that much data ;-)
No meaning beyond proximity, though? I've been looking for ways to "map" similar data so it will be (mostly) time-stable as nodes are removed and added, to no avail yet.
Posted by: Hans Gerwitz | July 11, 2006 at 09:05 AM
Hans,
Time stability is an interesting problem. One way to do this is to fix nodes in location
(or certain nodes). Alternatively, you could allow nodes to become more lethargic in movement
according to how long they have been there. This seems like a good idea. Are you going for
some form of animated representation? I'd love to hear more about it.
As for performance - I'm using pretty standard hardware, but I've spent a reasonable amount of time tuning the graph representation to make it faster. Many force based iterative layout implementations render the graph at each iteration - I simply iterate to the end and then display. This is an obvious point, but makes it a lot faster.
Posted by: Matthew Hurst | July 11, 2006 at 09:36 AM
Very cool! I found 3 of my blogs.
Posted by: Randy Charles Morin | July 13, 2006 at 11:54 AM
Hello Matthew,
Hello Hans,
You're both working on very interesting projects. Recently I started my own attempt in visualizing the blogosphere as an animated real-time application. If you're interested have a look here:
http://www.metaportaldermedienpolemik.net/wiki/Blog/2006-10-18/Animated+Visualization+Draft+of+the+Blogosphere
Kind regards,
Walter
Posted by: Walter Rafelsberger | October 26, 2006 at 10:38 AM
Thank you for the great article. Greetings from Germany.
Posted by: kostenlos Versicherungsvergleich | March 19, 2007 at 04:45 AM
great ones. keep them coming
Posted by: butter | February 06, 2008 at 04:38 PM