> From: Noelle <http://dummy.us.eu.org/noelleg> > Date: Sun, 9 Jun 2013 19:48:08 -0700 (PDT) > > she's probably wrong about this? Most definitely. It's extremely difficult *not* to be able to identify people, even with a highly anonymized data set. For example, when Netflix released their anonymized data so people could improve their recommendation algorithm, algorithms were eventually able to identify every person in that data set. > > Date: Mon, 10 Jun 2013 02:41:31 +0000 > > From: "Susan Reimers (Twitter)" < > > > > Susan Reimers @Meezer_Mom > > @Noellegi I've no issue w/data aggregated for trend purposes so long as > > personal identifiers not retained. THAT is an issue. > > 02:41 AM - 10 Jun 13