The Library of Congress is Archiving 170 Billion Tweets — on Tape
BY Julia Wetherell | Monday, January 7 2013
When the Library of Congress teamed up with Twitter in 2010 to archive four years’ worth of activity on the microblogging platform, the aim was to preserve a slice of early-millennial life to future researchers. Now the two-hundred-year-old institution is grappling with the resulting 133 terabytes of data, a bundle that includes every 140-character message sent out from Twitter’s six-year-history, from its inception in spring 2006 to December 2012.
While the initial agreement called for archiving up to 2010, the LOC subsequently determined to extend the project indefinitely, keeping up with the nearly half-billion tweets dispatched every day. There will be a six-month holdout before new tweets enter the archive, an interesting statement on our contemporary definition of “history.” The struggle now is for the LOC to create a keyword-searchable catalog for the vast amount of metadata associated with the archive, including the time and location that indicate a tie to certain events — election night on 2008, as one example. However, as the LOC’s recent report confirms, their agreement with Twitter states that the “Library cannot provide a substantial portion of the collection on its web site in a form that can be easily downloaded.” Therefore, it goes on to say, the archive will exist primarily in the physical realm – on tape.
The technical infrastructure for the Library’s Twitter archive follows the same general practices for monitoring and managing other digital collection data at the Library. Tape archives are the Library’s standard for preservation and long-term storage. Files are copied to two tape archives in geographically different locations as a preservation and security measure.
While this storage method seems oddly similar to creating nuclear-winter-proof seed vaults, it does anticipate the fact that, with all of our cloud-tending tech, we don’t exactly have a permanent record for our online life. For that, future historians may have the Library of Congress to thank.
The headline of this post has been corrected. An earlier version implied the Library of Congress report said Twitter posts were being stored on an analog media, which it does not say.