Comments on: Week 10 Lifestream Summary https://edc15.education.ed.ac.uk/champton/2015/03/21/week-10-lifestream-summary/ Just another Education and digital culture 2015 site Mon, 30 Mar 2015 20:36:39 +0000 hourly 1 https://wordpress.org/?v=4.2.38 By: PJ Fameli https://edc15.education.ed.ac.uk/champton/2015/03/21/week-10-lifestream-summary/#comment-380 Fri, 27 Mar 2015 14:01:07 +0000 http://edc15.education.ed.ac.uk/champton/?p=557#comment-380 Clare, good summary of Tweetorial and concluding paragraph. I share your perspective about quality vs. quantity of Tweets. Cheers, PJ

]]> By: Clare https://edc15.education.ed.ac.uk/champton/2015/03/21/week-10-lifestream-summary/#comment-379 Fri, 27 Mar 2015 03:44:48 +0000 http://edc15.education.ed.ac.uk/champton/?p=557#comment-379 Rather disconcerting that it can be that far out – and that it’s trying to determine gender based on inference rather than fact.

]]> By: sbayne https://edc15.education.ed.ac.uk/champton/2015/03/21/week-10-lifestream-summary/#comment-378 Thu, 26 Mar 2015 16:02:23 +0000 http://edc15.education.ed.ac.uk/champton/?p=557#comment-378 That’s interesting Katherine – I’ve been digging around a bit for more on how these things infer gender from social media – it sounds like it’s pretty sketchy and an emergent field using things like first names categorisation, linkage to weblog profiles and text/image processing, so we’ve clearly surfaced something not quite working with Keyhole et al here… Shiny, convincing new truths must be being made all the time by algorithms that only partly work.

]]> By: jdarling https://edc15.education.ed.ac.uk/champton/2015/03/21/week-10-lifestream-summary/#comment-377 Tue, 24 Mar 2015 15:59:26 +0000 http://edc15.education.ed.ac.uk/champton/?p=557#comment-377 >> after more careful analysis I’m not sure that I was any the wiser for it

I came to the same conclusion. Looks good but doesn’t deliver anything new.

]]> By: Katherine https://edc15.education.ed.ac.uk/champton/2015/03/21/week-10-lifestream-summary/#comment-375 Sun, 22 Mar 2015 03:03:49 +0000 http://edc15.education.ed.ac.uk/champton/?p=557#comment-375 Hi Claire, I went and did some counting between the Keyhole and Tweet Archivist records. (Assuming both were using the same data…) Actually, the contribution from female participants is more like 36-46% (depending on number of participants or number of tweets). That is nothing like 10%! I don’t know how Keyhole is judging gender (Twitter doesn’t require it to sign up for a new account), but it’s clearly undercounting!

]]>