/* ---- Google Analytics Code Below */

Tuesday, October 28, 2008

Visual Analysis Needs are Examined

Was hoping to make it to the InfoVis conference last week in Columbus, but could not make it work. I see that Bryan Pierce writes about the conference in a post: Are visual analysis tools poised to become pervasive?, which covers Christian Chabot's talk re flawed assumptions about why people use visual analysis. (looking for a video of the talk) Good perceptive read. These learnings are good and I generally agree with them ... most people's problems have data, complexity, and needs that can be solved on small machines today ... given the right tools.

Based on my own experience with analysis problems, there are often cases where a person without modeling training 'hits the wall' and cannot move forward without help. Visualization, though always useful, may not be enough to step forward and solve some problems. There is a temptation then to over-simplify to be able to show the pictures. The cost of error can be high. You may want to have a professional available for help.

So I agree, visual analysis can solve many problems .... people should be trained in easy to use tools. They should also be first trained in 'stating the problem' in such a way that it clearly and correctly states a goal. That part at least should be collaboratively reviewed, based on the cost, opportunity or otherwise, of solving the wrong problem, or getting the wrong solution for the sake of a picture. Saw that when spreadsheet tools became common, and seeing some of that now.

No comments: