Title: Building workflows for near-term iterative forecasting
Organizers: Gallagher, Alyse Cara, Bektaş, Billur, Singh, Patrícia, and Paniw, Maria
https://euro-ecoforecast.wordpress.com/
Duration: 9-12 a.m.
Content
If you work with individual monitoring data of plants and animals, you may have asked: How can we build forecasts that evolve and improve as new data arrives? In this hands-on demonstration, we will guide participants in developing near-term iterative forecasting workflows using individual-level plant monitoring data. Using worked examples in R, we will demonstrate how to prepare and use these data, apply different modelling frameworks (e.g., Joint Species Distribution Models), quantify uncertainty, and update forecasts as new response and explanatory variables become available. Rather than emphasizing any single modelling approach, we will focus on building flexible, dynamic, and reproducible workflows that can accommodate multiple methods and improve over time.
Participants will gain practical experience using R, Git, and GitHub to design automated pipelines for forecasting. We will also emphasize tools for visualizing predictions and communicating uncertainty—key components for engaging stakeholders effectively. As part of this, we plan to include a short section focused on data visualization principles, including practical before-and-after examples.
As a highlight, participants will have the opportunity to engage with one of Europe’s first biodiversity forecasting challenges, submitting and evaluating their forecasts based on a real-world case study. We will conclude by outlining future directions for improving spatiotemporal biodiversity predictions through better workflows, uncertainty communication, and stakeholder integration. Basic knowledge of R and a GitHub account are required to fully participate.
