I heard a talk by Ben Shneiderman about information visualisation yesterday for the Cambridge Usability Professionals Group. (It was ironic that I had a “locational usability” problem and was almost late, having made the novice error of trying to find Microsoft Research in the William Gates Building – which is indeed named after Microsoft’s Bill Gates foundation – but Microsoft Research in Cambridge was set up by Roger Needham, so it is in his building!)
The talk itself was very easy to follow with lively demonstrations of a number of visualisation tools. Shneiderman was very careful to point out that you need to have a good question and good data to get good results from information visualisation, and that it is no panacea, but when it works, it is fantastically powerful. One of the key strengths is that it makes it easy to spot outliers or anomalies in huge masses of data, particularly when there is a general underlying correlation. It is almost impossible to detect trends in a big spreadsheet full or numbers, but convert that into a visual form and the trends leap out. This means that you can see at a glance things like which companies’ stocks are rising when all the others are falling. Of course, graphs are nothing new, but the range of analytical tools that are now available mean you can quickly pick out things like spikes and shapes in your data in a way that would have been painstaking previously. There are also very important applications in medical research and diagnosis, as the ability to track which order certain events happen, helps researchers establish whether one condition causes another and could even be used to generate personal health alerts.
I liked the smart-money style treemaps (although the choice of red-green can’t be great for anyone who is colourblind), but I found the marumushi newsmap fun but not much more informative than traditional newspapers, mainly because the newsmap crams in more words than I can take in. Newspapers are really pretty good at writing headlines that work, and you can usually see at a glance what today’s top story is anyway – it’s the one in big letters at the top! However, if you need an aggregation of global news for international comparison, the newsmap does give you quick access to a lot of international sources all together.**
One of the great pleasures of these events is getting to talk to other people who are there and I met a fascinating researcher who had been monitoring importance of stories by keyword frequency, showing that when something happens you get a burst of news activity around the relevant keywords, a ripple effect, and then it dies away. By looking at those patterns you can produce a measure of the impact of different events.