BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20220812T074334Z
LOCATION:Samarkand Room
DTSTART;TZID=Europe/Stockholm:20220627T133000
DTEND;TZID=Europe/Stockholm:20220627T140000
UID:submissions.pasc-conference.org_PASC22_sess129_msa161@linklings.com
SUMMARY:Socio-Epidemiological Insights from a Yearlong COVID-19 Twitter St
 ream
DESCRIPTION:Minisymposium\n\nSocio-Epidemiological Insights from a Yearlon
 g COVID-19 Twitter Stream\n\nMassemin, Gupta, Hemmatirad, Raileanu, Müller
 ...\n\n<br />The social media infodemic of COVID-19 had dynamics comparabl
 e to viral transmission. This study explores a yearlong Twitter data strea
 m and quantifies the value of social media trends as proxies of epidemiolo
 gical events and policy effectiveness. We collected 793 million tweets bet
 ween throughout 2020, and filtered for multilingual COVID-targeted keyword
 s in URLs or tweet text across more than 20 languages. Topics are generate
 d as groups of hashtags using Word2Vec embeddings on tweet text and cluste
 red using a novel variant of DBScan (DBScan-v), which forms clusters of va
 riable density and number. We explored correlations of these topics, meta-
 topics and multilingual sentiment analysis (using LASER embeddings) with g
 lobal epidemiological trends, such as the number of SARS-CoV-2 tests, COVI
 D case counts, viral transmissibility, mobility, and policy stringency. Se
 veral strong (Pearson coefficient > 0.9) correlations were found, such as 
 transit trends negatively correlated to lockdown topics (e.g. #workfromhom
 e), and policy stringency positively correlated with topics describing eco
 nomic hardship and mental health issues. Statistical analysis of social me
 dia trends shows its potential use in predicting epidemiological events an
 d it may stimulate socially-aware policymaking. Finally, the generated top
 ic clusters can serve as independent datasets or validate keyword filters 
 for further analyses.\n\nDomain: Computer Science and Applied Mathematics,
  Engineering
END:VEVENT
END:VCALENDAR