By Andrew Johnson, postbacc at North Carolina State University
Folks came in from across the country earlier this week to contribute to SAMSI’s Multivariate Modeling in Ecology workshop. Living in Raleigh, I shared in the excitement but enjoyed a much shorter commute than most. During my drive to SAMSI on the morning of the first day, I must say that I had a few reservations. I was new to the working group and unsure about my ability to contribute to the group’s progress. I was immediately relieved to find that my fears were misplaced. The atmosphere throughout the workshop was warm and industrious, welcoming of all questions and suggestions. Two of my favorite features were the flexibility of the schedule and the fluidity of the subgroups, both being obvious products of our group’s openness to input. There was a predetermined schedule for each day that was quickly adjusted by the group’s suggestions. This was also the case for group sizes. When appropriate, the group would quickly split into smaller, project-based subgroups. These self-guided adaptations made it possible for everyone involved to get the most out of their experience.
The workshop was designed to address a few of the issues that the Multivariate Modeling working group has grappled with in the past. The group has been interested in understanding how the spatial positions of multiple species can be interrelated and co-vary with environmental influences, as well as the best way to model those relationships. Needless to say this poses a few inherent challenges. One of the primary questions that the subgroups focused on was that of dimension reduction for datasets describing the interactions among multiple species and environmental variables. Complementing the efforts on dimensional reduction, the subgroups also worked on developing improved methods for interpreting their computational results. In addition to the questions above, our group discussed strategies for assembling separate datasets into one coherent “picture.” This required consideration of each dataset’s reliability and corresponded to the computational weight placed on each. Fortunately, my favorite question of “where to eat” was answered with ease each day, as the workshop was fully catered with excellent food!
I gained a great deal from this workshop, and had a blast doing it! I am thrilled to have been a part of it.