Dec 18, 2009 (11:12 AM EST)
Full Nelson: Many Eyes, Dancing Data
Read the Original Article at InformationWeek
Data's got a tough year ahead of it. It will be, at once, transparent, open, and luminous. It will captivate in real time, but also predict the future. It will fix the banks and car companies, but also tear down reputations and brands in litigation. It will get breached, kidnapped, and, in some cases, remain untapped. Its ubiquity -- Excel, database queries, dashboards -- will cause blindness in the cubicle and heart attacks in the boardroom. Data needs more tools like IBM's Many Eyes, an "alpha works" Web-based visualization tool that combines insight, creativity, and collaboration.
It's not a big revelation that data will factor into most every major decision a corporation makes. For example, during the past year, bankers, analysts, and arm-chair pundits have laid the blame for economic collapse at the feet of technology, citing lack of proper visibility into real-time data, weaknesses in predictive modeling, and disconnected silos of data . . . all of which is a bit like Tiger Woods blaming his BlackBerry.
Nevertheless, nearly every financial services CIO is talking about data - more, better, faster, more tightly integrated. Andy Wood, the CIO at Wilton Re, a re-insurance company, says business intelligence is the only application his company custom built.
Government leaders, all the way up to the tippy top, are talking about making data open and transparent, whatever that means. Retailers are using data to continually transform their supply chains and re-imagine internal and partner processes. Healthcare professionals are standardizing on electronic medical records. And so on.
2010: The year of data.
Many Eyes is exciting, then, because it takes data out of the weeds and the wonk and breathes in some real-world life. It began in 2007 out of IBM Research Labs in Cambridge, Mass.; the name comes from the open source saying "Many eyes make bugs shallow." There are similar products from Swivel and Data360, each with impressive data sets (Many Eyes has more than 3,000 at this count), but it's the actual visualization that separates Many Eyes.
(You can watch a video ReviewCam of Many Eyes by clicking on the "Play" button directly below.)
For example, one data set tracks Twitter activity on 1,000 stocks and visually presents this data in a bubble chart, where the size of the bubbles represents the relative number of mentions, reflecting the Twitter buzz. Or view a Treemap of auto mileage data, but sort that data by type of car (large, compact, midsized) or make or transmission type. Or use a Scatterplot to show the relationships between any of these data points.
IBM makes Many Eyes fairly easy to use. Anyone can provide a data set with a simple copy and past in a form field (more information about data formats is available here), and then create visualizations. Drop down selections make it easy to pick from charting types. It's also easy to re-arrange and re-cast the data in a new light with simple selections and trial and error.
Anyone can create visualizations from the existing data sets, and then share them. Like this look at European Defense R&D spending:
One of the more curious aspects is the use of free form text, which can be formatted into bubble charts and word trees, where the exploration gets more interesting. For example, some visualizations analyze the relationships between words in a speech, or provide insights into what people are saying about a topic (or person) on Twitter.
Here are a couple of examples -- trends in business intelligence (just to be relevant) and the relationship of verbs and nouns in A Christmas Carol (just to be timely):
For now, one of the downsides of this alpha works project is that it's public, but customers can work with IBM to create private views of data, on a personal or business basis. As data takes center stage in 2010, it wouldn't be surprising to find more companies taking advantage of this "visualization as a service" project.