Which field should I look for in a UMM-C record to find the Principal Investigator (or science team member responsible) for a given data collection?  Is it different from model to model?

  • No labels

13 Comments

  1. That would be a ContactPerson. Contact Persons in UMM-C can have the following roles: "Data Center Contact", "Technical Contact", "Science Contact", "Investigator", "Metadata Author", "User Services", "Science Software Development".

    Contact Persons can be a top-level element on the collection and can also come under the DataCenters.

  2. Not all P.I.s want to be contactable though.  Is there a Citation field where they would be cited as the author of the dataset?

  3. Also, using a Contact field would be problematic for an investigator that is retired or deceased; yet the product should still be findable by the P.I. name and citable with the P.I. as author.

  4. Erich Reiter Can you please help answer some of these questions? 

    Christopher Lynnes As for citations, we are implementing those this PI, so no collections translated from XML formats will have citations populated yet. Erich can answer if the model we will be implementing will cover your use case.

    1. But if it is the metadata anywhere, I should be able to see it in the Native form in CMR, yes?

  5. Here is an example of a collection in the UMM specification :

        "CollectionCitations": [{
            "Creator": "Collection Citations Creators1",
            "Editor": "Collection Citation Editors1",
            "Title": "Collection Citiation Title1",
            "SeriesName": "Series Name 1",
            "ReleaseDate": "2017-01-01T12:00:00Z",
            "ReleasePlace": "New York, NY",
            "Publisher": "Some Publishing Co.",
            "Version": "1",
            "IssueIdentification": "12345",
            "DataPresentationForm": "Word",
            "OtherCitationDetails": "Other Citation Details",
            "OnlineResource" : {
                "Linkage" : "http://www.google.com?q=publishers",
                "Protocol" : "http",
                "ApplicationProfile" : "Microsoft Word",
                "Name" : "Collection Citations Online Resource 1",
                "Description" : "Collection Citations Online Resource 1",
                "Function" : "information"
            }
        },
        {
            "Creator": "Collection Citations Creators2",
            "Editor": "Collection Citation Editors2",
            "Title": "Collection Citiation Title2",
            "SeriesName": "Series Name 2",
            "ReleaseDate": "2017-01-01T13:00:00Z",
            "ReleasePlace": "New York, NY",
            "Publisher": "Some Publishing Co.",
            "Version": "2",
            "IssueIdentification": "12345",
            "DataPresentationForm": "Word",
            "OtherCitationDetails": "Other Citation Details",
            "OnlineResource" : {
                "Linkage" : "http://www.google.com?q=publishers",
                "Protocol" : "http",
                "ApplicationProfile" : "Microsoft Word",
                "Name" : "Collection Citations Online Resource 1",
                "Description" : "Collection Citations Online Resource 1",
                "Function" : "information"
            }
        }],
    
    


    The creator is the collection author and should be the principle investigator.  DIF 9 and 10 both support this element.  ECHO 10 however does not.  The citation is just one long string and while the investigator should be in that string, we can't parse it out.  It will be translated though into the OtherCitationDetails in the UMM.  So if the PI is in the citation you can see their name if you get the record in its native format.  In ISO SMAP and MENDS you will find it by looking in the identification section in the citation section under cited responsible party with a role of author. 

    .../gmd:identificationInfo/gmd:MD_DataIdentification/gmd:citation/gmd:CI_Citation/gmd:citedResponsibleParty

    While we have the citation information in the metadata if the provider provides it, The CMR/EDSC doesn't have a way to search for a collection by collection author or contact person.  This is a capability that would need to be added.

    1. Is the Citation information at least indexed? If that is the case, then a free text search ought to at least pick it up.

      1. No it is not.  Currently the CMR doesn't have the capability to search citations, collection authors, nor metadata or collection contacts.

        1. That is in the plan for this PI and we can expect it to be done by the end of this PI. Currently, we have tickets to index and search the author. If we need additional fields indexed and searchable, let me know and I will write the tickets for those.

          1. The difficult specification is ECHO 10 where the entire citation is in a string and we will have difficulty parsing it.  We could have 2 indices. One for the author and 2 for the concatenation of all of the citation elements. The question that I have is from EDSC how would a user expect to search for authors do we search both and give more weight to the author index?

            1. Strictly speaking, parsing is not necessary to achieve the desired effect (in most cases). If it is treated as just a free text search, like Description/Abstract, then a search on PI name will find the record most of the time, because PI last names tend to stand out from the rest of the text.  (Except for Suomi. But that's what you get when they name a mission after you.)

              1. True - so you would be OK if we indexed DIF 9 and 10 creators (collection authors), ECHO 10 the entire citation, and ISO SMAP/MENDS individual names into 1 index and search on that.  That should be pretty easy to do. - Chris' reply is to this paragraph only.

                Or we could combine all of the citation fields together into one sum and index that for each specification. Add a separate author search string as well as add it to the keyword search. 

                Sorry I am just rambling and probably shouldn't have posted this (smile) 

                1. Yes, that would be OK.