Ecosystem modeling with presence-only data

I had the pleasure of hearing a presentation by Nick Record and Ben Tupper of the Bigelow Laboratory on how to forecast things like tick and jellyfish levels, when all or nearly all of the data you have were reported by volunteers, making it “presence-only” data. Volunteers could, in principle, report the absence of something, but they seldom do. You would think this would be next to impossible, given that some areas have no one reporting, others have a single volunteer reporting faithfully, and others have occasional reports whenever someone thinks of it. And yet you are nevertheless able to turn this low-quality evidence into something useable. Fascinating stuff. Learn more here.