Soil, Streams and Rivers: integrating Data Science and Biogeochemistry using Inductive-Deductive approaches
|
Figure by Kristen Underwood
|
The patterns and processes approach is based on the two main types of reasoning we use in data science versus biogeochemistry or related disciplines. The figure on the left shows how these approaches can be combined: inductive approaches begin with observing a pattern, in data sciences typically by exploring data, while deductive approaches begin with testable hypotheses. Wherever we start, there likely will be steps of refinement, be it about scales, research questions, or the generalizability of findings.
In my work with the Big Data cluster and my research more broadly, I enjoy working with data scientists and complex systems scientists that support process ideas with pattern analyses and develop algorithms based on questions. These folks include Kristen Underwood, Donna Rizzo, Byung Lee (all UVM), Scott Hamshaw and Dustin Kincaid (USGS), Li Li and Bryn Stewart (Penn State). |
