Experimental Design

I have learned during my career that many experimental investigations were inefficient. These were accompanied by a poor understanding by the technical teams of the power of statistical analysis and a similar level of understanding by the statistical teams of the underlying physical basis of the experimental factors being studied. I have a strong skill in designing experiments using statistical DOE techniques that take into account the realistic interactions that will occur on a physical basis and then translate those into a robust and frugal design. For example, I successfully reduced a response surface study for an analytical method that would have required more than 500 manual assays (equivalent to more than 4 FTE weeks) to less than 50 assays.

DOE Statistics
JMP Excel (complex macros)
Large datasets Datamining
Visualization Pattern finding

Important to designing, implementing and analyzing efficient and robust experiments are the experience and expertise of appropriate software tools. My two primary tools are Microsoft Excel and SAS Institute’s JMP. I have written simple and complex Excel macros to achieve goals such as automating retrieving and analyzing data, as well as directly controlling hardware. Where more powerful analysis is required I use JMP. I also use it to create highly visual reports and presentations. For large, confounded and/or sparse datasets I have an innate ability to quickly find patterns that are not readily seen by others. I was the first person to use JMP within product development at GSK and it has become a tool used by the majority of scientists. I apply my experimental design skills across all of my areas of focus described elsewhere.

Download a PDF information pack describing background experience and expertise, and areas of technical focus.

Information pack

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.