While March came in like a lion weather-wise in many places, the saying meant good news index-wise for ArchiveGrid. This week we updated the index and once again, the number of collections and items represented in ArchiveGrid spiked. In January’s update, the index grew by around 600,000 records and reached 2.4 million. Now we’re over 3 million.
MARC records from WorldCat represent at least 90 percent of ArchiveGrid’s descriptions. Since no particular MARC record field tells us “Hey! Include me in ArchiveGrid,” we use a combination of elements. In the “recall vs. precision” performance metric, we’ve tended to err on the side of recall. Details on how we filter WorldCat records for inclusion in ArchiveGrid are here.
Twice this year, we tuned the filter we use to extract MARC records from WorldCat to include more record types based on the MARC Leader byte 6 value. January’s update brought in records with the value of “k” (two-dimensional, non-projecting graphics). This update includes records with the value of “g” (projected medium), “i” (nonmusical sound recording), or “j” (musical sound recording). We think these adjustments allow more descriptions of the types of materials ArchiveGrid searchers could expect to find, without overloading the index with records we’d prefer to filter out: irrelevant materials or published works, items held in multiple locations, etc. So we expect to continue adjusting the filter and see the total number of records change as we get more precise.
Here are some highlights of what valuable primary sources the added g, i, and i indexes in ArchiveGrid have to offer:
- Around 1,800 sound discs, tape reels, and cassettes of nearly all of Duke Ellington’s commercial and non-commercial recordings and also some radio broadcasts. Collected by Joseph Jeffers Dodge, Harvard University acquired the collection in March 1998.
- A live 1963 recording in Germany of John F. Kennedy’s “Ich bin ein Berliner” speech, at Ball State University.
- A May 2, 2003 VHS recording of the “Service of death and resurrection for Fred McFeely Rogers, or Mister Rogers, at Pittsburgh Theological Seminary.
Another addition with this update is documentation for ArchiveGrid indexes. We use a number of “hidden” indexes in ArchiveGrid for testing and trouble-shooting, so in a new how to search page, we explain what these indexes are and how they can be used in a search. This should be considered a work in progress, so if you have suggestions for improvements or questions about how the indexes work, please let us know.