Discovering data

Discovery (Note: this case is still being developed)

How to determine which dataset you want based on several similar looking collections

• distinguishing between different types of datasets
  • descriptions aren’t always helpful
  • crawling catalog links, automated harvesting
  • determining the popularity of the data for research
  • “yelp” for collections, “this data was useful” or “this data is crappy”
  • maybe use: http://schema.org/Review

Using DOIs for bidirectional links to research/publications and back to originating data

• Interesting Questions:
  • What has this data been used for in research?
  • What data was used for this research, can I reproduce this work?
• Dryad (http://wiki.datadryad.org/Main_Page):
  • it involves submitting actual data, not pointers to the data
• Including DOI creation and citation as part of the process for data (not just publications)
  • how to enforce this?
  • incentives for data producers vs providers
  • feeds back into the rating system