expedition clustering
Using unsupervised learning techniques to recreate expedition clusters from archival museum collections data
The California Academy of Sciences holds records for over a million botanical specimens, but it’s often unclear which collecting expedition each one came from. The Expedition Clustering project groups individual specimens back into their expeditions, which makes the collection easier to organize, summarize, and explore.
The approach clusters specimens by where and when they were collected and who collected them. Specimens gathered close together in space and time, by the same people, most likely came from the same trip. Knowing this helps researchers trace historical and modern expeditions, highlights regions that have been under-sampled, and supports maps and visualizations for storytelling and teaching.
I worked on this as a volunteer with the Academy’s Scientific Computing Department from 2024 to 2026.