I'm planning to dive into SAS JMP (pronounced "jump") for the next few weeks and figure out the ins and outs of serious data visualization. Up until now, I've done most of my visualizations in Excel, which is quite useful, but rather limiting for more advanced visuallizations and large data sets. As a first step in this process I reviewed the SAS webinar, "Visualization of Data and Models," by John Sall, co-founder and Executive Vice President of SAS. In this slightly geeky overview of the JMP visualization capabilities Sall walks the audience through a variety of different visualization scenarios: Napoleon's march on Russia; Hurricane patterns; crime stats; selecting the optimal percentage of material to create a tire tread; etc..

More over, he actually manipulates the data live as he is presenting. If you have ever tried to do live data manipulation during a presentation, you will appreciate how difficult this can be! However, Sall is trying to reenforce the idea that data is best presented and manpulated right in JMP without using a slide deck. This is a nice idea, if only I could figure out how to get my IT department to install JMP on our conference room computers... However, for we lesser folks with limited JMP access, the tool does offer Flash exports of "some" visualizations.

The following are some of my key takeaways and screen captures from the presentation.

Sall notes that you can get the same linear regression for wildly different sets of data and we humans are quite excellent at detecting visual patterns, so creating a graph is a handy way to quickly see what's going on with your data.

Sall also shows off some of the new features in JMP 9 where data can be overlayed on top of interactive maps. He uses the data from Napoleon's march on Russia and monthly hurricane data to illustrate this cool new feature.

Napleon lost 396,000 of his original 400,000 troops during his march on Russia--a staggering 99% of his army! The following graph is presented by JMP as a full motion visualization where you can see Napoleon's army diminishing in size.

Sall also shows a full motion visualization of hurricanes moving across the Atlantic and increasing and decreasing in size and frequency over time. It's a pretty great resource for meteorologist trying to understand and predict weather patterns.

Finally, I found the three dimensional visualizations of optimization problems quite useful, as one can imagine applying this functionality to all sorts of things. As a bonus, the optimization visualizations produce very pretty graphs. In the following visualization Sall is showing how one might go about using JMP to select the best percentage of ingredients to include in a formula to produce optimal tire treads.

In conclusion, SAS JMP is a very valueable tools for creating all sorts of visualizations. Sall give a good high level overview of JMP's capabilities. I would be hesitant to do live visualizations for my presentations, but when you are a co-founder of SAS you can do what you want.

For those interested, SAS also provides the following links for futher exploraton:

• Review questions & answers from webinar.
• Read a white paper by Chuck Pirrello of JMP for more data visualization best practices (PDF).
• Share this white paper by Jon Weisz of JMP with others in your organization to help them understand different types of models and the value they can bring to your organization (PDF).
• Attend a Training Course on visualization using JMP by attending JMP Software: Data Exploration . View the entire JMP learning path for a complete view of what is available and stay tuned for more developments in the form of new courses and more coverage of JMP 9 features.
• Get the new SAS Press book JMP Essentials: An Illustrated Step-by-Step Guide for New Users  by Curt Hinrichs and Chuck Boiler.