Posts Tagged ‘texts’

mining for big history: uncouth things i want to do with archives

Wednesday, June 24th, 2009

woohoo THATcampers!  i’m so psyched to hang out with you.  actually, i need to learn from your enormous brains…

a major theme of my graduate course in digital history at the u of c was the opportunities lying around unprecedented scale of access and manipulability.

historians, for instance, typically train to write 20- to 40-year studies, at most 100-year histories; they frequently teach by the century, at most the five-hundred-year time period.  proposal: digital archives, as a revolution in access, radically open the horizons for legitimate big history of long-term trends.

ideas for sessions:

* how would you text mine a 500-yr history?  how bout a 5000-yr history?  many of the tools for text-mining (cf philologic) look narrower and narrower within a peculiar text; how could these tools be used to crunch many texts across large time periods (off the top of my head: graph for me, computer, the top verbs used around the word “eye” in medical texts since Greece …  )?  how can timelines more usefully render the results visual (and interactive!)?

* how bout images.  here we’re talking about 200 years   what can you do with 1 billion photographs?  what happens when you automagically photosynth (livelabs.com/photosynth/) the entire nineteenth- to twentieth-century city of London?  what about “averaging” photos: www.faceresearch.org/demos/average ?  what does the average house look like, decade by decade?  what does an average coal miner look like?

* how bout maps.  doug knox (hi doug!) and i have been talking with the newberry map librarians about how you’d collate atlases of place names, travelers’ diaries, and maps to annotate an interactive atlas of chicago where any given block could be peeled back, year by year.  how would you make a 300-year thick map of the american west?

Here's what others are saying about THATCamp on Twitter