Life in the network_ the coming age of computation

Lifeinthenetwork:thecomingageofcomputationalsocialscienceDavidLazer,HarvardUniversityAlex(Sandy)Pentland,MITLadaAdamic,UniversityofMichiganSinanAral,NYUAlbertLaszloBarabasi,NortheasternUniversityDevonBrewer,InterdisciplinaryScientificResearchNicholasChristakis,HarvardUniversityNoshirContractor,NorthwesternUniversityJamesFowler,UCSDMyronGutmann,UniversityofMichiganTonyJebara,ColumbiaUniversityGaryKing,HarvardUniversityMichaelMacy,CornellUniversityDebRoy,andMITMarshallVanAlstyneBostonUniversityWelivelifeinthenetwork.Whenwewakeupinthemorning,wecheckoure-mail,makeaquickphonecall,walkoutside(ourmovementscapturedbyahighdefinitionvideocamera),getonthebus(swipingourRFIDmasstransitcards)ordrive(usingatranspondertozipthroughthetolls).Wearriveattheairport,makingsuretopurchaseasandwichwithacreditcardbeforeboardingtheplane,andcheckourBlackBerriesshortlybeforetakeoff.Orwevisitthedoctororthecarmechanic,generatingdigitalrecordsofwhatourmedicalorautomotiveproblemsare.Wepostblogentriesconfidingtotheworldourthoughtsandfeelings,ormaintainpersonalNIHPublicAccessAuthorManuscriptScience.Authormanuscript;availableinPMC2009September16.Publishedinfinaleditedformas:Science.2009February6;323(5915):721–723.doi:10.1126/science.1167742.NIH-PAAuthorManuscriptNIH-PAAuthorManuscriptNIH-PAAuthorManuscriptsocialnetworkprofilesrevealingourfriendshipsandourtastes.Eachofthesetransactionsleavesdigitalbreadcrumbswhich,whenpulledtogether,offerincreasinglycomprehensivepicturesofbothindividualsandgroups,withthepotentialoftransformingourunderstandingofourlives,organizations,andsocietiesinafashionthatwasbarelyconceivablejustafewyearsago.Thecapacitytocollectandanalyzemassiveamountsofdatahasunambiguouslytransformedsuchfieldsasbiologyandphysics.Theemergenceofsuchadata-driven“computationalsocialscience”hasbeenmuchslower,largelyspearheadedbyafewintrepidcomputerscientists,physicists,andsocialscientists.Ifoneweretolookattheleadingdisciplinaryjournalsineconomics,sociology,andpoliticalscience,therewouldbeminimalevidenceofanemergingcomputationalsocialscienceengagedinquantitativemodelingofthesenewkindsofdigitaltraces.However,computationalsocialscienceisoccurring,andonalargescale,inplaceslikeGoogle,Yahoo,andtheNationalSecurityAgency.Computationalsocialsciencecouldeasilybecomethealmostexclusivedomainofprivatecompaniesandgovernmentagencies.Alternatively,theremightemergea“DeadSeaScrolls”model,withaprivilegedsetofacademicresearcherssittingonprivatedatafromwhichtheyproducepapersthatcannotbecritiquedorreplicated.Neitherscenariowillservethelong-termpublicinterestintheaccumulation,verification,anddisseminationofknowledge.Whatpotentialvaluemightacomputationalsocialscience,basedinanopenacademicenvironment,offersociety,throughanenhancedunderstandingofindividualsandcollectives?Whataretheobstaclesthatstandinthewayofacomputationalsocialscience?FromindividualstosocietiesTodatethevastmajorityofexistingresearchonhumaninteractionshasreliedonone-shotself-reporteddataonrelationships.Newtechnologies,suchasvideosurveillance,e-mail,and‘smart’namebadgesofferaremarkable,second-by-secondpictureofinteractionsoverextendedperiodsoftime,providinginformationaboutboththestructureandcontentofrelationships.Considerexamplesofdatacollectioninthisareaandofthequestionstheymightaddress:Videorecordingandanalysisofthefirsttwoyearsofachild’slife(1)Preciselywhatkindofinteractionswithothersunderliesthedevelopmentoflanguage?Whatmightbeearlyindicatorsofautism?Examinationofgroupinteractionsthroughe-maildataWhatarethetemporaldynamicsofhumancommunications—thatis,doworkgroupsreachastasiswithlittlechange,ordotheydramaticallychangeovertime(2,3)?Whatinteractionpatternspredicthighlyproductivegroupsandindividuals?Canthediversityofnewsandcontentwereceivepredictourpowerorperformance(4)?Examinationofface-to-facegroupinteractionsovertimeusingsociometersSmallelectronicspackages(‘sociometers’)wornlikeastandardIDbadgecancapturephysicalproximity,location,movement,andotherfacetsofindividualbehaviorandcollectiveinteractions.Whatarepatternsofproximityandcommunicationwithinanorganization,andwhatflowpatternsareassociatedwithhighperformanceattheindividualandgrouplevels(5)?Lazeretal.Page2Science.Authormanuscript;availableinPMC2009September16.NIH-PAAuthorManuscriptNIH-PAAuthorManuscriptNIH-PAAuthorManuscriptMacrocommunicationpatternsPhonecompanieshaverecordsofcallpatternsamongtheircustomersextendingovermultipleyears,ande-CommerceportalssuchasGoogleandYahoocollectinstantmessagingdataonglobalcommunication.Dothesedatapaintacomprehensivepictureofsocietal-levelcommunicationpatterns?Whatdoesthe“macro”socialnetworkofsocietylooklike(6),andhowdoesitevolveovertime?Inwhatwaysdotheseinteractionsaffecteconomicproductivityorpublichealth?TrackingmovementWithGPSandrelatedtechnologies,itisincreasinglyeasytotrackthemovementsofpeople(7,8).Mobilephones,inparticular,allowthelargescaletracingofpeople’smovementsandphysicalproximitiesovertime(9),whereitmaybepossibletoinferevencognitiverelationships,suchasfriendship,fromobservedbehavior(10).Howmightapathogen