BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Every tweet counts: How statistical content analysis of social net
 works can improve our knowledge of citizens' preferences. An application t
 o France and US presidential elections and EU leaders' popularity - Prof. 
 Stefano Iacus\, Dipartimento di Economia\, Management e Metodi Quantitativ
 i Universita' di Milano
DTSTART:20121121T141500Z
DTEND:20121121T151500Z
UID:TALK41194@talks.cam.ac.uk
CONTACT:Stephen Clark
DESCRIPTION:The growing usage of internet and social media by a wider audi
 ence of citizens sharply increases the possibility to investigate the web 
 as a device to explore and track their (policy) preferences\nand judgement
 s. For the statistician social networks data are interesting for at least 
 two reasons: users of social networks express mainly unsolicited opinions 
 and the volume of these opinions is extremely\nhigh\, i.e. this is the big
  data world. In the present talk we present the statistical ideas behind t
 he so called Hopkins and King's method for content analysis as well as sev
 eral successful applications to\npolitics and political competitions. As s
 imple case studies\, we present the results obtained by "Voices from the B
 logs" (http://voicesfromtheblogs.com\, a research project developed at the
  University of\nMilan) on the 2012 Presidential ballot and in the subseque
 nt Legislative election\, the recent US Presidential competition and EU le
 aders popularity. Despite internet users are not necessarily representativ
 e\nof the whole population of country's citizens\, our analyses show a rem
 arkable ability of social-media to forecast electoral results as well as a
  consistent correlation between social-media results and the\nones obtaine
 d in more traditional mass surveys. In some cases\, the possibility of an 
 'information overload' arises as a factor affecting negatively the analysi
 s of social-media.
LOCATION:Lecture Theatre 1\, Computer Laboratory
END:VEVENT
END:VCALENDAR