RTP JUG kicked off 2016 right! Two dynamic speakers generated engaging discussions and ideas exchanges during the RTP JUG meeting on January 12 at the SAS campus in Cary, NC. Buffy Hudson-Curtis, a consulting non-clinical statistician from GSK discussed an analysis of a DOE experiment, and Michelle Woolfolk, an engineer from the City of Durham Stormwater Services, presented complex data used to protect and manage the Falls Lake reservoir.
The DOE case discussion described an analysis of a tablet dissolution study which examined the effects of several process parameters and formulations. Buffy Hudson-Curtis offered a structured method used to develop her experiment, moving from visualizing the results with the JMP variability platform, to the analysis in Fit Model, through construction of the final figure and finally development of specific wording to describe her results. The data set generated many interesting lines of discussion that will likely lead to further talks on split plot designs and experimental plans designed to reduce variation. In the case study, one of the variables had three levels could have been coded as a nominal or continuous variable (or even a discrete numeric). Should the variable be modeled as categorical or with a quadratic term? Buffy walked the group through the differences and explained why it really depends on the question being asked by the investigator, and more importantly, the question being asked by the customer/management.
Got Water? Thank Michelle Woolfolk and her team; it’s not that way by accident. North Carolina's relatively flat topology means that we have shallow lakes (actually man-made reservoirs.) Managing these natural resources and the water entering them is a complex task. Michelle described how she monitors environmental sample data to make decisions about how to manage inputs from water treatment plants, storm water run-off, agricultural land and mother nature's constant barrage. The best reactive and proactive plans to respond to changes in the input stream rely on her ability to understand the current state and trends in the Falls Lake biome. Michelle walked the group through JMP visualizations designed to help her and her customers understand the metrics that are critical to sustained water quality where we live. The results and implications from the complex battery of water tests without clear visual communication would be muddy at best.
There were two action items at the end of the meeting:
1. We need speakers for the upcoming meetings.
2. We need 2-3 more people for the leadership team.
Please send me a note to let me know your topic and availability.
It sounds like there is a lot of interest in talking more about how to analyze different types of DOE and similar data sets. Post your ideas for topics that you think would be interesting.