Usage-based Data Discovery - 1st meeting

From Federation of Earth Science Information Partners

Slide Deck for Meeting

Participants

  • Chris Lynnes (EOSDIS) - Lead
  • Bob Downs (SEDAC): Intrigued by user perspective, looking to broaden discovery to include the exploration needed before deciding on datasets
  • Carson Davis (NASA/IMPACT): Working on Knowledge Graph, corpus of 90K research articles related to Earth Observations
  • Doug Newman (Data Use Architect for EOSDIS EED-2 contract): working with Chris Lynnes on EOSDIS knowledge graph
  • Ed Armstrong (PO.DAAC): long term interest in discovery. Worked on AIST MUDRODS project, which included a discovery component
  • Joe Lee (The HDF Group): works on HDF use website
  • Jonathan Blythe (BOEM): works on Environmental Studies. Creating a KnowledgeBase, wants to see that go beyond individual systems
  • Karen Moe (Disasters Cluster): works with people who are not data experts, looking for easier way to find the data they need
  • Mark Parsons (RDA): has worked on data citations for a long time. Researchers also want to know who is using their data
  • Megan Carter (ESIP Community Director): provide ESIP help. N.B.: Call for summer sessions coming soon!
  • Siri Jodha Singh Khalsa (NSIDC, RDA, GEO): Chair of RDA Data Discovery Interest Group, also working on Geo Knowledge Hub (the inspiration for the EOSDIS Knowledge Graph!)
  • Tyler Stevens IEOSDIS): worked on GCMD, now on CMR
  • Dave Neufeld (NOAA)

Key Points

  1. It's important to record why/how we know about a given relationship, not just because someone says so. (Esp. pertinent for the more general entities like Sustainable Development Goals). Must be able to do objective analysis on the graph.
  2. The W3C Provenance model may be helpful here. Also, nano-publication.
  3. How to handle subsets, e.g., spatio-temporal subsets?
  4. Should consider whether tool relationships are within scope.
  5. Precipitation datasets seem like a good starting point.

Actions

  1. Lynnes - write up some text for Megan's newsletter
  2. Lynnes - set up regular meeting with ESIP G2M, some time other than 4th Thursday
  3. TBD - spin off a subgroup to work on use cases
  4. Lynnes/Newman - generate a sample export of EOSDIS knowledge graph
  5. Blythe - provide info on use of W3C Prov model