This SESUG conference was a very energetic and rewarding experience, with many user presentations on graphics using SAS/GRAPH and ODS Graphics.
One standout presentation for me was on "Data Merging and Exploration to Identify Associations Between Environmental Factors and Disease Outbreaks" by Neeta Shenvi, et. al. The key part of interest to me and readers of this blog was the extensive and sophisticated usage of GTL to create various graphs. Neeta and her associates used every trick in the book, including a lattice layout, common uniform axes, eval() functions and all this with macro %do loops to populate the number of plots in a cell dynamically.
On my part, I presented the following seminars and presentations:
- Clinical Graphs using SAS - a 1/2 day pre-conference seminar.
- Hands-on presentation on ODS Graphics Designer.
- Super Demo on New SAS 9.3 Features in SG Procedures.
- Keynote presentation on "Make a Good Graph".
The ODS Graphics Designer continues to gather more interest, as many users want to use an interactive application to create graphs. With Designer, you can create graphs without typing a single line of code. The Designer generates GTL code behind the scenes, and you can view this code as you go. This code can be copied to the program editor for execution. Using Designer is also a great way to learn GTL.
New SAS 9.3 features for SG procedures such as cluster groups for all plot types, and support for interval axis for Box Plot were very popular with the attendees. These features make it easy to create clinical graphs. One user commented that now he can replace pages of code with these simple options. See this previous article for more details.
The opening session included a debate on the importance of Data v/s Analytics by Maura Stokes and Diane Hatcher. This session was hilarious, with Diane emphasizing the importance of data and Maura extolling the virtues of analytics. Through use of humor, they succeeded in providing valuable information.
For the keynote presentation, I presented the techniques useful to create graphs that are effective in communicating the information easily and clearly with minimum distractions. I discussed the application of some findings in the field of visual perception, and well established recommendations from thought leaders for optimizing data ink and eliminating chart junk. I also presented examples on how to reduce eye movement required to decode a graph, direct display of data and use of proximity for comparisons.
With over 450 attendees, SESUG 2012 was indeed a very successful regional SAS user's conference.