[Metadatalibrarians] Normalization solutions for GIS keywords?

[Joyce Chapman]

We are looking into normalizing incoming FGDC keyword metadata to improve our local GIS data search. We are surveying the terrain to see what other folks are up to in terms of dealing with the problem of uncontrolled incoming keywords assigned by metadata creators. There are a few options as we see it: leave as is and let users browse/search uncontrolled metadata provider keywords (what we currently do), normalize the keywords as they come in using a controlled vocabulary, retain existing keywords but assign additional categories on top of providers' keywords to serve as a browsable grouping for users (seems like many use ISO 19115 Topic Categories for this), or retain keywords but also assign additional keywords from some something like LCSH. We've recently conducted a study with advanced GIS researchers and found that leaving the metadata provider assigned keywords for search and browsing can be quite confusing for them, leading to incomplete search results, missed resources, and general confusion.

We were wondering if people wouldn't mind sharing their own experiences with us.

* What processes did you consider to fix the problem of uncontrolled keywords in incoming FGDC metadata and how they affect browsing and searching within your systems for the users? 

* What kind of staff time has your solution required? 

* Did you consider normalizing incoming keywords, and why didn't you if not? 

* What vocabularies are you using or did you consider using, and why? 

* Does anyone have before/after assessment data that implies an improved user experience created by your efforts? 

* What metadata fields are you indexing in a user's keyword search? 

 Thanks so much,
 Joyce Chapman, Liz Bell, and Jeff Essic



Joyce Chapman
Libraries Fellow 
North Carolina State University
joyce_chapman at ncsu.edu
919.513.1652