ESPM 288: Reproducible and Collaborative Data Science

Ecological Forecasting Challenges


Format

This semester (SPRING 2021) will be entirely different. Students will work in self-assembled teams to design and enter the international ecological forecasting challenge. We will not have synchronous meetings or recorded lectures; I will meet (remotely via Zoom) with each team on it’s own schedule to help them design and implement their forecasts. Students will also interact with and learn from the broader ecological forecasting community participating in the challenge.

Overview

This spring I will be offering a modified version of my graduate seminar, ESPM-288: “Reproducible & Collaborative Data Science”, which will focus on more advanced topics in data science applications, in particular, on Ecological Forecasting.

The NSF funded EFI Research Coordination Network (EFI-RCN) has just launched the Ecological Forecast Challenge with the goal to create a community of practice that builds capacity for ecological forecasting by leveraging National Ecological Observatory Network (NEON) data products. Over the course of the semester, student teams will select and enter into one or more of the five challenge areas (Aquatic Ecosystems, Terrestrial Carbon and Water Fluxes, Tick Populations, Phenology, Beetle Communities).

The course will be entirely remote and primarily asynchronous. Instruction will be tailored to individual team needs, but will generally focus on computational, informatics, and data science aspects of building ecological forecasting scripts which can download data, run forecasts, and publish results autonomously, as well as an overview of core statistical concepts (e.g. “strictly proper” scoring rules for forecasts). However, our emphasis will be on student-led investigation and development of forecasts. Students will be expected to form their own teams and explore, develop, and test their own forecasts at their own pace.

All team submissions will also be entered into the international EFI NEON challenge. This will be the only graded assignment for the seminar, and should be of suitable publication quality work. Entrants into the challenge will likely also join a co-authored manuscript on the final challenge results.

Similar seminars focused on entering the challenge are also being run at a handful of other institutions around the country. Students have the opportunity to interact not only with other graduate seminars, but also with research labs around the country that will be entering the challenge. More details about the challenge are available at: https://ecoforecast.org/efi-rcn-forecast-challenges/. If you’ve taken ESPM-288 before but would still be interested in participating in the challenge, please just let me know.