Isn't it great to be in Market Research? For now, all of you who are reporting quantitative data to your clients can, and should start exploring these graphic possibilities. From today, consider me a "Many Eyes Junky"....
Thursday, 31 January 2008
Graphing the impossible: Data Visualisation
Hans Donkers of the Dutch market research company Stratus shared his view on the future of market research at an SPSS Seminar today. It was good to hear how he and I agree on some of the mayor trends today that undoubtedly will reshape the future of our industry. After defining the term Research 2.0, he addressed the five new technologies that may reshape our industry. He distinguished analysing structured data and ways to analyse unstructured data. Obviously the industry needs to find its way to be able to consistently analyse blogs, tags, social networks, and instant messages. In fact, my post below addresses analysis of tags. In another post last year, I addressed our industry's shift away from measuring claimed behaviour to measuring actual behaviour. Today, Donkers was underpinning the fact that more and more data will become available (at notionally no costs). His example of google having 2.7 billion hits a month on average in 2007 was striking. I was pleasantly surprised to see how he than shared current examples visualising the impossible: How to graphically visualise 2.5 million consumer ratings of 9276 music artists? It seems I've been missing the entire trend. Was I too busy preparing ordinary PowerPoint slides to present to my clients? There's a whole world out there that will undoubtedly change the way we'll present our data to clients. I am most impressed by a platform that was introduced a year ago: IBM's Many Eyes. Users can upload any data set, and the tools for visualizing and graphing the data are incredibly richer. The visualisation options include country maps, line, stack, pie, and bar charts, block histograms, bubble diagrams, scatter plots, network diagrams, a lot! Another great example is one by the Stanford University. Together with Yahoo! Research Labs, and MIT. The paper is here. The authors developed a set of interactive visualisation tools and used them on 2.5 million Yahoo! user ratings of 9,276 music artists. The computes layout below shows the grouping of artists and helps for browsing large music collections. How else could we ever conclude that Mel C = Britney Spears = Nick Carter?