What if users could find Earth science data by starting with how they plan to use it? For instance: "I want to predict landslides: which datasets will help me do that?" This is the problem that the Discovery Cluster is currently addressing. This session will bring together coders, data and information curators, and end users to develop simple prototypes to explore ways in which intended data usage can lead directly to finding the most useful datasets. Bring your own use case or work on one of the Cluster's use cases. Bring your own usage-data relationships, or work with the Cluster's datasets.
PLEASE
SIGN UP HERE FOR ONE OF THREE TEAMS BEFORE (so we can optimize your Hackfest experience)
Also, the detailed agenda is online
HERE.
View Recording
View Session Notes
View Presentations: See Session Notes above
Takeaways- Three successful implementations of usage-based discovery indicates the paradigm is feasible (though scalability requires further work).
- UX work showed that the most appropriate target user may not be the end-user of an application, but rather the data/application wrangler that supplies the end user.
- Manual foraging for application-dataset relationships was hard: better practices of data citation by applications would make this easier and more tractable to automation.
