Ecology, statistics, and programming enthusiast
Broadly, I am studying how we can use data as effectively as possible. In ecology, data are expensive and time-consuming to collect, so it’s important to consider what we can learn from the data we have, and to prioritize what data to collect in the future.
Here are some specific questions I have been exploring:
Some species have an inordinate effect on the rest of the community, and losing these species can result in ecosystem collapse. These species are of interest both to ecologists and for identifying conservation priorities. However, identifying these species using a traditional field approach is time-consuming and often not feasible. I’m looking at several potential correlates of species importance, including population variability and network centrality measures.
Most current ecological networks incorporate only a single type of interaction data (consumer-resource interactions, for food webs, or mutualistic interactions, often plant-mutualist, for mutualistic networks). We know these interaction types co-occur, but what can we learn by including multiple interaction types in a single network? I have used three ecological networks to study how adding or removing interaction types changes our understanding of the roles species play in their communities. This study was published in PLoS Computational Biology.
The border image at the top of the screen is actually adapted from this project. It’s an alluvial diagram comparing species role in a complete community to a community without herbivores.
It is possible to fit a mechanistic model to data when studying a small closed system, say, with one or two species. But this approach gets increasingly difficult and data-hungry when considering more species. I am currently studying how we can use Dynamic Bayesian Networks and Causal Bayesian Networks to understand species interactions in complex ecological communities.