3 ms·

The data is here: http://wikileaks.org/gifiles/ http://wikileaks.org/gifiles/

• olefoo 15y ago
  That's not particularly useful to people who want to rip through the entire corpus using standard textmining tools; we'd really rather a signed torrent of a compressed imap archive or something similar. Having all the headers in formats accessible to scripts to read and tag different emails, etc. This is sort of like having to do research on an archive that's been photocopied onto kleenex. But, we'll suck it up, and write a scraper to turn it into what we need. But it means that some of the interesting stuff is delayed, because wikileaks is trying to get the most bang for the buck by dribbling it out. Just to pick an example of what I mean by interesting; tag all emails that mention a country or city and see if there are correlations between who sent and received emails and the countries mentioned, or if informants are mentioned in connection with more than one country.