Comments on: Weekend Minis – Design Paradigms, Colbert Bump, and Bullet Graphs http://flowingdata.com/2008/02/09/weekend-minis-design-paradigms-colbert-bump-and-bullet-graphs/ Strength in Numbers Sun, 12 Feb 2012 03:48:24 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Hadley http://flowingdata.com/2008/02/09/weekend-minis-design-paradigms-colbert-bump-and-bullet-graphs/#comment-49800 Hadley Sat, 09 Feb 2008 14:53:00 +0000 http://flowingdata.com/2008/02/09/weekend-minis-design-paradigms-colbert-bump-and-bullet-graphs/#comment-49800 I like the "there is no single view" article, but I think it neglects to very important tools for understanding data - interaction and models. Interaction gives us linked brushing, which is a particularly flexible tool for comparing different subsets of the data. Models give us the capability to describe and remove strong trends in the data so we can see subtler effects that were previously masked. Tukey described this process as using residuals and re-iteration. I also wonder about the use of other people's software for all the bad examples, and their own for all the good examples. Can you do unbiased research while making money off of your software? I like the “there is no single view” article, but I think it neglects to very important tools for understanding data – interaction and models. Interaction gives us linked brushing, which is a particularly flexible tool for comparing different subsets of the data. Models give us the capability to describe and remove strong trends in the data so we can see subtler effects that were previously masked. Tukey described this process as using residuals and re-iteration.

I also wonder about the use of other people’s software for all the bad examples, and their own for all the good examples. Can you do unbiased research while making money off of your software?

]]>