Data Science: Research directions
- Updated
On this page
Keeping meaning and provenance through reuse
Emerging direction · Research and infrastructure direction; reviewed 2026-11-03
Combining datasets requires consistent descriptions of what was measured and how it may be used. Machine-readable metadata can help, but incompatible concepts and incomplete documentation still create work. Better discoverability must be accompanied by provenance and justified interpretation.
The FAIR Guiding Principles — GO FAIR Foundation · Checked
Communicating the limits of a result
Emerging direction · Research direction; reviewed 2026-11-03
Uncertainty methods can communicate more than a single predicted value. Their assumptions must remain visible when an analysis is reused or conditions change. A numerical interval should explain what it covers rather than becoming a general seal of certainty.
A Gentle Introduction to Conformal Prediction — Angelopoulos and Bates / arXiv · Source date: 2021-07-15 · Checked