Digifeet15 Seminar, December 2nd 2015 Seminar program Preparing and analyzing large social media datasets #Digifeet15
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Digifeet15)Seminar,)December)2nd)2015)
Seminar)program)))
)
)Preparing)and)analyzing)large)social)media)datasets))
#Digifeet15)
Table&of&contents&
Seminar(Participant(List………………………………………………………………………………………….......2(
Seminar(Program……………………………………………………………………………………………………......3(
&Abstracts&
Anders(Søgard(”Language(Technology(and(Social(Media”…………………………………………….5(
Hilke(Reckman(”Towards(modeling(writing(and(sociolects(in(social(media(–(capturing(writing(styles(with(characterElevel(ngram(models”………………………….…………………………..6(
Kristoffer(Laigaard(Nielbo(”Brothers(in(Arms(–(Outlining(of(a(Jihadology(Macroscope”(7(
Per(Møldrup(Dalum(”Big(Data(Meets(the(Danish(Cultural(Heritage”………………………......8(
Luca(Rossi(“Dynamic(socioEtechnical(platoforms,(limits(and(possibilities(for(social(media(data(analysis”………………………………………………………….......................................................9(
Anja(Bechman(”Do(you(people(post(less(on(Facebook?(Challenges(of(Big(Data(methods”…………………………………………………………………………………………………………………..10(
Ravi(Vatrapu(”Social(Set(Analysis:(A(SetETheoretical(Approcah(to(Computational(Social(Science”…………………………………………………………………………………………………………………….(11(
Anders(Kristian(Munk(”Remediating(Instagram(for(Obesity(Research?”……………………..12(
Jiyoung(Kim(“Twitter(and(Big(Data(Methods………………………………………………………………13(
Jakob(Linaa(Jensen(”Mapping(news(on(social(media”…………………………………………………15(
Guide:(How(to(find(Dokk1(and(Hærværk(Restaurant………………………………………………….16(
2
Seminar(participants(
(Anders'Geil,'Aarhus'University'([email protected])''Anders'Kristian'Munk,'Aalborg'University'Copenhagen'([email protected])''Anders'Søgaard,'University'of'Copenhagen'([email protected])''Anja'Bechmann,'Aarhus'University'([email protected])''Hilke'Reckman,'MIT'Media'Lab,'Massachusetts'([email protected])''Jakob'Linaa'Jensen,'Danish'School'of'Journalism'([email protected])''Jiyoung'Kim,'Yeungnam'University,'South'Korea'([email protected])''Kristoffer'Laigaard'Nielbo,'Aarhus'University'([email protected])''Leon'Derczynski,'University'of'Sheffield,'UK'([email protected])''Luca'Rossi,'IT'University'([email protected])''Mette'Simonsen'Abildgaard,'Aalborg'University'Copenhagen'([email protected])''Morten'Krogh'Petersen,'Aalborg'University'Copenhagen'([email protected])''Per'Møldrup'Dalum,'State'Library'Aarhus'([email protected])''Peter'B.'Vahlstrup,'Aarhus'University'([email protected])''Ravi'Vatrapu,'Copenhagen'Business'School'([email protected])''Thomas'Bjørnsten,'Aarhus'University'([email protected])'
(
! 3!
Seminar(Program(The$seminar$location$is$Dokk1,$Hack$Kampmanns$Plads$2,$8000$Aarhus$C,$Meeting$Room$1.$
The$participants$each$have$20$minutes$to$do$their$presentation.$In$the$end$of$each$session,$there$will$be$time$for$a$short$discussion$on$the$topics.$The$moderators$will$control$the$time$and$initiate$and$manage$session$discussions.$$$Wednesday,(December(2,(10.00817.00($10.00G10.15:$Welcome$
Preprocessing(of(Social(Media(Data(Leon(Derczynski(as(moderator 10.20G10.40:$Anders$Søgaard,$University$of$Copenhagen$
10.45G11.05:$Hilke$Reckman,$MIT$Media$Lab,$Massachusetts$$
11.10G11.20:$Short$discussion$
11.20G11.35:$Short$break$
Machine(Learning(and(Big(Data(Leon(Derczynski(as(moderator((11.35G11.55:$Kristoffer$Laigaard$Nielbo,$Aarhus$University$
12.00G12.20:$Per$Møldrup$Dalum,$State$Library$
12.25G12.35:$Short$discussion$
12.35G13.35:$Lunch$
Social(Media(and(Methods(Anders(Geil(as(moderator( 13.35G13.55:$Luca$Rossi,$IT$University$
14.00G14.20:$Anja$Bechmann,$Aarhus$University$
14.25G14.45:$Ravi$Vatrapu,$Copenhagen$Business$School$
#Digifeet15(
! 4!
14.50G15.00:$Short$discussion$
15:00G15.15:$Short$break$$!Social(Media(Cases(Thomas(Bjørnsten(as(moderator!!
15.15G15.35:$Anders$Kristian$Munk,$Aalborg$University$Copenhagen$$
15.40G16.00:$Jiyoung$Kim,$Yeungnam$University,$South$Korea$$
16.05G16.25:$Jakob$Linaa$Jensen,$Danish$School$of$Journalism$
16.25G16.40:$Short$discussion$
$
16.45G17.00:$Concluding$Remarks$by$Leon$Derczynski$$17.45:$Dinner$at$Hærværk$Restaurant,$Frederiks$Allé$105,$8000$Aarhus$
5
Keynote'Abstracts'
Language'Technology'and'Social'Media!
!
Anders'Søgaard,'center'for'Language'Technology,'University'of'Copenhagen!
!
What!do!Twitter!users!think!of!Lars!Løkke!Rasmussen?!Where!on!Facebook!should!you!
place!an!ad!for!home!contents!insurances?!Who!is!the!best!candidate!for!a!specific!job!
on!LinkedIn?!The!bottleneck!to!solve!these!kinds!of!tasks!is!to!understand!what!people!
write!on!the!social!media.!Language!technology!(NLP)!uses!machine!learning!to!
develop!tools!to!automatically!analyze!a!text.!In!this!presentation,!the!challenges!
related!to!automatic!analysis!of!a!text!on!social!media!will!be!reviewed.!Some!things!
are!much!harder!than!usual,!some!much!easier.!I!will!try!to!explain!why!–!and!go!
through!examples!of!research!in!improving!of!language!technology!for!social!media.!!
6
Towards(modeling(writing(styles(and(sociolects(in(social(media(–(capturing(writing(styles(with(character6level(ngram(models(
Hilke(Reckman,(MIT(Media(Lab((
In#the#case#of#online#product#reviews,#the#information#available#about#the#author#of#a#
review#is#often#limited#to#a#user#name#and#the#review#itself.#However,#for#the#company#
that#sells#the#product#it#is#interesting#to#know#how#it#is#received#by#different#
demographic#groups.#Character=level#ngram#models#have#been#argued#to#be#a#good#
tool#for#guessing#properties#like#gender#and#age#of#authors.#I#present#a#small#test#of#the#
applicability#of#this#technology#to#short#online#product#reviews,#as#well#as#to#tweets.#
Then#I#propose#clustering#similar#authors#based#on#their#writing#styles,#using#such#
models.#The#idea#is#to#see#which#groups#emerge#naturally,#rather#than#using#
predefined#groups#in#terms#of#the#traditional#demographic#categories#such#as#gender#
and#age.#The#interests#and#preferences#of#such#groups#can#then#be#modeled,#using#
profile#information#where#available,#as#well#as#topic#extraction,#sentiment#analysis,#etc.#
(
(
(
7
Brothers(in(Arms(–(Outlining(of(a(Jihadology(Macroscope(
Kristoffer(Laigaard(Nielbo,(Aarhus(University(
The$link$between$Islam$and$violence$has$become$a$central$topic$in$public$debate$due$
to$elaborate$and$often$aggressive$internet$strategies$from$Islamist$organizations$such$
as$Al;Qaeda,$AQAP$and$IS.$A$widespread$position$in$this$debate$argues$that$Islamic$
scriptures$in$and$of$themselves$serve$as$primary$motivators$for$violence.$To$investigate$
this$claim$we$extract$and$model$posts$and$articles$from$jihadology.net$(an$online$
database$for$jihadi$primary$source$materiel).$This$talk$will$present$a$model$of$AQAP’s$
Inspire$magazine,$in$which$we$use$Latent$Dirichlet$allocation$to$probe$similarities$
between$discourses$related$to$violence,$religion$and$family.$$
$
8
Big$Data$Meets$the$Danish$Cultural$Heritage$
Per$Møldrup6Dalum,$State$Library,$Aarhus$
DeIC%(Danish%e,Infrastructure%Cooperation)%and%the%State%and%University%Library%have%
agreed%to%establish%the%DeIC%National%Cultural%Heritage%Cluster,%State%and%University%
Library.%The%cultural%heritage%cluster%applies%state,of,the,art%technologies%within%data%
science,%and%for%the%first%time%ever%facilitates%quantitative%research%projects%on%the%
digital%Danish%cultural%heritage%–%e.g.%radio%and%TV%programmes,%the%Danish%Web%
Archive,%and%historical%newspapers.%
The%Cultural%Heritage%Cluster%is%making%IBM's%BigInsights%platform%available%to%research%
projects.%This%platform%consists%of%the%Open%Data%Platform%(ODPi)%and%number%of%
commercial%applications:%BigSheets,%BigSQL,%BigR%and%Text%Analytics.%Combined,%these%
four%systems%form%the%basis%for%carrying%out%analyses%by%means%of%known%techniques%
on%very%large%data%sets.%
$
9
Dynamic(socio+technical(platforms,(limits(and(possibilities(for(social(media(data(analysis(
Luca(Rossi,(IT(University,(Denmark(
Social'media'have'been'part'of'our'lives'for'more'than'a'decade.'Even'if'they'are'often'
still'perceived'as'something'“new”'they'have'been'through'several'evolutions'and'
transformations.'Even'when'a'single'platform'has'survived,'new'features'have'been'
implemented'or'pulled'down,'new'social'practices'have'emerged'and'new'marketing'
strategies'have'been'adopted'producing'social'media'platforms'that'are'profoundly'
different'from'when'they'started.'This'short'presentation'will'explore'the'
consequences'on'digital'data'production'of'this'unstable'nature'of'social'media'
platforms.'It'will'explore'how'different'social'practices'will'result'in'different'largely'
incomparable'data'and'how'this'affects'the'possibility'of'long'term'longitudinal'
studies.'Presenting'open'problems'from'ongoing'researches'the'presentation'will'
discuss'a'possible'strategy'to'deal'with'this'problem.''
(
10
Do#young#people#post#less#on#Facebook?#Challenges#of#Big#Data#methods#
Anja#Bechmann,#Aarhus#University#
The$talk$will$present$a$regression$study$of$the$development$in$the$number$of$posts$on$
a$national$dataset$of$1000$Danes’$private$Facebook$data$from$2007@2014$and$discuss$if$
young$people$post$less$in$the$last$period,$thereby$supporting$the$idea$that$young$
people$post$less$on$Facebook$and$instead$use$Instagram$and$Snapchat.$Furthermore$
the$presentation$will$discuss$if$the$Pareto$principle$apply$to$Facebook$posting$patterns.$
Lastly$the$presentation$will$draw$on$work@in$progress$on$the$mapping$of$topics$in$
Facebook$groups$in$Denmark$and$Korea$and$discuss$challenges.$The$research$question$
is$what$kind$of$patterns$occur$in$the$secret,$open$and$closed$groups$in$the$two$
countries.$Key$issues$are$what$is$the$population,$how$do$we$recruit$and$what$are$the$
challenges$in$normalizing$and$mining$social$media$data$and$can$we$trust$such$mining$
techniques.$$
11
Social'Set'Analysis:'A'Set/Theoretical'Approach'to'Computational'Social'Science'
Ravi'Vatrapu,'Copenhagen'Business'School,'Denmark'
Current'research'projects'at'the'Computational'Social'Science'Laboratory'
(http://cssl.cbs.dk),'Copenhagen'Business'School'are'addressing'an'academic'research'
gap'and'real@world'industry'need'to'describe,'model,'analyse'and'explain'large@scale'
interactions'on'organisations’'social'media'channels'as'individuals''associations'to'
ideas,'values,'identities'etc.'Towards'this'end,'we'are'developing'and'evaluating'a'set@
theoretical'approach'to'big'data'analytics'termed'“Social'Set'Analysis”'(SSA).'Social'Set'
Analysis'consists'of'three'primary'research'activities:'(a)'theorising,'modelling,'and'
collecting'big'social'data'about'organisations'(e.g.,'Danish'Cancer'Society’s'official'
Facebook'page);'(b)'combining'those'big'social'data'sets'with'in@house'organisational'
data'sets'(e.g.,'Customer'Relationship'Management'systems);'and'finally'(c)'analysing'
the'combined'datasets'by'applying'set'theoretical'methods'and'tools'(crisp'sets,'fuzzy'
sets,'rough'sets,'random'sets'and'Bayesian'sets).'This'talk'will'outline'the'SSA'
approach,'report'selected'empirical'findings,'discuss'implications'and'limitations,'
identify'challenges'and'future'research'directions'
12
Remediating*Instagram*for*Obesity*Research?*
Anders*Kristian*Munk,*Aalborg*University,*Denmark*CoAauthors:*Morten*Krogh*Petersen*&*Mette*Simonsen*Abildgaard.***
The$concept$of$obesogenic$environments$has$gained$traction$in$the$broader$field$of$
obesity$research$within$recent$years.$It$captures$an$as$yet$vaguely$defined$idea$that$
obesity$is$related$to$complex$everyday$live$contexts.$Notions$like$sociality$and$
uncertainty$may$thus$be$crucial$to$understanding$what$drives$obesity.$In$a$recent$
experiment$with$a$trans=disciplinary$group$of$obesity$researchers$and$cultural$analysts$
we$conducted$a$series$of$experiments$on$how$to$cut$large$amounts$large$amounts$of$
Instagram$data$in$a$ways$that$would$be$amenable$to$the$study$of$obesogenic$
environments.$These$experiments$were$predicated$on$the$idea$that$the$everyday$is$on$
Instagram$(or$perhaps$that$Instagram$is$part$of$the$everyday)$and$that$the$platform$
thus$affords$opportunities$for$a$cultural$analysis$of$obesity.$Teasing$out$these$everyday$
environments$from$tens$of$thousands$of$geotagged$images$turned$out$to$pose$
interesting$research$problems$in$and$of$itself.$$
*
13
Twitter'and'Big'Data'methods'
'
Jiyoung'Kim,'Yuengnam'University,'South'Korea'
'
Jiyoung(Kim(will(present(two(studies(that(include(Twitter(and(big(data(methods(such(as(
webmetrics(and(social(network(analysis.(The(first(study(provides(an(empirical(and(
longitudinal(account(of(how(the(South(Korean(political(podcast(series(Nakkomsu((NKS),(
an(alternative(media(outlet,(employed(a(hybrid(media(strategy(during(three(elections(
in(South(Korea(between(December(2011(and(December(2012.(The(study(is(based(on(
the(methodological(triangulation(of(hybrid(web(indicators(resulting(from(a(social(
network(analysis,(a(semantic(network(analysis,(and(a(link(impact(analysis,(all(
conducted(using(big(data(mined(from(social(media(during(the(13Fmonth(period.((
(
To(examine(the(emergence(and(evolution(of(communication(patterns(around(NKS(on(
Twitter,(all(tweets(that(were(publicly(accessible(and(contained(the(Korean(word(for(
NKS(were(collected.(As(a(result,(a(data(set(of(79,028(unique(vertices((users)(and(
1,866,085(edges(was(obtained.(Our(findings(include(the(following.(First,(network(
density(increased(gradually(over(the(period,(reflecting(a(continuous(decrease(in(the(
number(of(users(participating(in(discussions(around(NKS.(Second,(according(to(
quadratic(assignment(procedure((QAP)(correlations,(the(semantic(network(
characteristics(of(these(discussions(changed(over(time.(Third,(the(results(of(the(
semantic(network(analysis(and(the(link(impact(analysis(highlight(the(NKS(production(
team’s(hybrid(media(practice(as(well(as(the(intervention(facing(the(team.(In(this(
context,(Twitter(served(as(a(space(where(individual(listeners(engaged(in(direct(
interactions(with(podcast(producers(and(shared(relevant(information(with(fellow(
listeners,(that(is,(a(space(that(podcasts(themselves(did(not(provide.((
(
Another(noteworthy(characteristic(of(the(NKS(phenomenon(was(the(range(of(offline(
activities(in(which(listeners(were(encouraged(to(participate(alongside(the(podcast,(such(
14
as(books(by(panelists,(public(talks,(and(fundraising(events.(Such(activities(were(
characterized(not(only(by(their(critical(content(but(also(by(elements(of(playfulness.(The(
second(study(examines(how(cultural(innovation(on(Web2.0(using(Psy’s(Gangnam(Style(
effect(on(Twitter(by(considering(the(adoption(and(dissemination(of(Korean(popular(
music((Kpop)(on(Twitter.((The(data,(we(considered(Tweets(include(the(#Kpop(hashtag(
from(November,(2011,(to(February,(2013.((The(period(of(Tweets(are(divided(with(two(
periods(as(each(8months(before(and(after(Psy’s(“Gangnam(Style”(released(as(well(as(
the(Youtube(video(updated.(We(employed(webometrics(indicators(to(delineate(the(
change(of(the(#Kpop(Twitter(network(and(compare(the(changes(of(Kpop(
communication(networks.((Also,(this(research(analyzes(the(use(of(URLs(and(hashtags(
related(with(Kpop.(The(results(indicate(that(Twitter(users(across(the(world(used(Twitter(
as(not(only(communicate(within(country(also(communicate(with(the(other(countries(
through(free(online(video(sharing(or(radio(service.(Psy(effect(influences(on(the(
structure(of(the(Kpop(communication(and(adoption(on(Twitter(in(terms(of(cultural(
innovation.((A(contribution(of(this(study(is(to(provide(an(empirical(cultural(innovation(
of(Korean(Wave(using(network(analysis.('
15
Mapping'news'on'social'media''
Jakob'Linaa'Jensen,'Danish'School'of'Journalism'and'Media'
Social'media'now'play'a'major'role'in'the'news'circuit.'Stories'are'shared,'curated'and'
commented'upon'via'social'media,'not'at'least'Facebook,'Twitter'and'Reddit.'This'
study'examines'the'use'of'Twitter'for'news'dissemination'in'Denmark.'It'follows'the'
Twitter'accounts'of'25'Danish'media.'Network'and'Metrics'analyses'are'applied'to'
catch'the'overall'picture'of'news'dissemination.'Further,'more'inEdepth'analysis'
reveals'what'kind'of'stories'are'shared.'
Besides'the'substantial'conclusions,'the'presentation'will'also'discuss'methodological'
aspects'of'harvesting'Twitter'and'Facebook'data'on'news'dissemination,'using'DMI'
TCAT'and'Digital'Footprints.'
'
16
How$to$find$Dokk1$and$Hærværk$Restaurant$$ Dokk1%address:$$Hack$Kampmanns$Pl.$2$$8000$Aarhus$C,$Denmark.$$$$From%Aarhus%Central%Station,%the%yellow%city%busses%12,%13%and%15%stop%at%Dokk1%Europlads%as%well%as%the%blue%regional%busses%100,%103,%109,%17,%118,%120,%121,%122,%123%and%200.%%%In%order%to%be%at%Dokk1$by$9.45%on%Wednesday%2nd%get%on%the%yellow$city$bus$15$towards$Skejby$Nordlandsvej%at%9.37%from%Park%Allé/%Aarhus%Banegårdsplads.%%Get%off%at%stop$number$2:$Dokk1.$Europaplads%%(9.43).%%You%will%find%Dokk1%ahead%of%you.%%$If%you%would%like%to%be%at%Dokk1$by$9.37,%get%on%yellow$city$bus$13$towards$Vejlby$Nord%at%9.26%from%Park%Allé/Aarhus%Banegårdsplads.%%Get%off%at%stop$number$2:$Dokk1.$Europaplads%(9.37).%%You%will%find%Dokk1%ahead%of%you.%%$Guidance%to%yellow%city%bus%13%and%15%–%a%five%minute%walk%from%Aarhus%Central%Station:%When%standing%outside%the%entrance%of%Aarhus%Central%Station/Aarhus%Banegårdsplads%facing%7W11,%you%must%cross%the%street%and%take%a%left.%When%you%get%to%the%traffic%light%by%Baresso,%cross%it%and%take%a%right.%Walk%straight%up%the%street.%You%will%find%the%stop%for%bus%13%and%15%on%the%left%side%of%the%road.%%If%you%would%like%to%walk%from%Aarhus%Central%Station/Banegårdsplads,%you%can%do%that%also.%According%to%Google%Maps,%the%estimated%time%for%the%walk%is%ten%minutes.%When%standing%outside%the%entrance%to%the%Central%Station%facing%7W11,%take%a%right%and%walk%down%“Ny%Banegårdsgade”.%Follow%it%until%you%come%to%the%traffic%light%–%take%a%right%here%(on%Søndergade).%You%will%come%to%yet%another%traffic%light.%Cross%it%and%walk%straight%onto%Toldbodgade%(you%will%pass%the%restaurant%“Flammen”%shortly%after).%Follow%Toldbodgade%until%you%get%to%“Europlads”.%From%here,%you%might%be%able%to%see%Dokk1.%Take%a%right%on%Europaplads%and%cross%the%bridge%–%you%will%get%to%the%entrance%of%Dokk1.%%%Hærværk$Restaurant$address:%Frederiks%Alle%105%8000%Aarhus%C%%To%get%to%Hærværk%Restaurant%from$Dokk1.$Europaplads,%get%on%blue$regional$bus$200$
#Digifeet15$
17
towards$Skanderborg$Busterminal%(Møllegade)%at%17.25.%%Get%off$at$Frederiks$Allé/Søgade%(17.34)%and%walk%500%meter%straight.%You%will%find%Hærværk%Restaurant%on%the%left%side%of%the%road.%$%
Related Documents
