Jon Collins, a Freeform Dynamics colleague, kindly introduced me to Touchgraph, a beautiful and practical way to discover the patterns hidden inside data sets. The data can be in local or remote databases, spreadsheets, XML files and other formats.
To give a taste of the functionality, Touchgraph has provided a number of useful and free services. You can rummage Amazon books, music and movies or present the results of a Google search. To illustrate the principles, I asked it to find the links between my top 50 Facebook friends:
The size, shape and colour provide an at-a-glance interpretation of clusters, relationships and relative importance. The detail appears in a table on the left of the screen (not shown here) and you are provided with a number of selection, filtering and editing tools.
In the picture above, I selected Gapingvoid's Hugh Macleod to see what mutual 'friends' we have. My own set of friends is quite limited because I reject most requests for 'friendship'. Anyway, back to the picture. The green London and San Francisco blobs are networks. You just click on one to see who belongs to it.
A Google search is probably more interesting, not to mention less narcissistic. After deleting the obvious rubbish hits, you can see the patterns hidden in your search results. It's like clustering on steroids.
Here are the results for a Google search for '"online information conference" 2007'. (Plug. Plug.)
The isolated pink cluster is 'marketing opportunities' and the isolated mauve cluster is 'committee members'. I threw out a couple of irrelevant clusters - you're bound to get them with Google.
Enjoy. See you at the conference?
No comments:
Post a Comment