Perry Haaland
My current interests are in developing curricula and methods for teaching Data Science to statistics graduate students. Important data science projects are typically going to be approached by a team. So one important area for curriculum development is how to effectively teach Team Data Science; that is, how statisticians can work as team members and how teams can organize to effectively solve complex problems
Understanding and using best practices in a team setting is fundamental to the success of the team. I view best practices as a set of organizational practices that not only include coding, but also the thought processes and resulting analysis strategies. So in addition to advising students on how to organize their code, I also provide a framework for approaching complex problems. In general, I find that these skills need to be learned through practical experience and mentoring.
Some particular topics of interest to me are the use of GitHub for version control, the organization of folders and files (including standard naming conventions), best practices for organizing code, best practices for writing code, and best practices for analysis strategies. I generally recommend the use of the R software language for most applications, but the practices that I advocate apply equally well to Python and other languages.
Another interest for me is how teams are organized for success. In particular, this involves the important roles in a team and how team members can work together for success. Statistical and software engineering technologies are important, but the best analyst or best coder is not necessarily the person who will contribute most effectively to the success of the team. I believe, however, that students can learn these skills through experience.
I am applying my interest in teaching data science as I assist Professor Steve Marron with the statistical consulting class, STOR 765. I help teach the class as well as volunteering for one-on-one mentoring sessions. As part of that class I also advise students on the master’s projects. I look forward to continuing to mentor and educate the next generation of statisticians and data scientists.
Recently, we added Team Data Science experiences to the course, with great success. During the Spring semester, we hope to have a project of appropriate complexity and scope that requires team work but is also within the scope of a one semester project. Students in STOR765 can volunteer to participate on the team, and we provide a hands on experience in a variety of roles, best practices, coding, and problem solving.
I retired from Becton Dickinson (BD), a large medical technology company, in the fall of 2017. I was the first statistician hired by BD, in 1984. I was privileged to work on many interesting projects while maintaining my academic interests. I was recognized as a BD Fellow in Statistics in Experimental Design. The BD Fellow is the company’s highest scientific recognition. Over the years, I have worked closely with Professor Marron and advised several Ph.D. students in the department in addition to supporting them as interns at BD.
During my career in industry, I specialized in experimental design, statistical graphics, machine learning, assay calibration, bioinformatics, cloud computing, dashboards for decision support, and high dimension/low sample size data methods.
As part of my work, I developed a number of R packages that have appeared on CRAN and BioConductor. I also did extensive internal software development including robotics, databases, laboratory management, cloud based bioinformatics systems, etc.