key:
[ , ] Neyokuqungaqunga Algorithmic ingxaki nge Google Flu ekwizinga. Funda iphepha yi Lazer et al. (2014) , kwaye ubhale omfutshane, email ngokucacileyo njengenjineli Google echaza ingxaki nomnikelo ufifi ukulungisa le ngxaki.
[ ] Bollen, Mao, and Zeng (2011) uthi ukuba data evela Twitter ingasetyenziswa ukuqikelela imarike imfuyo. Esi siphumo kukhokelele ekuyilweni kwingxowa-watshintsha uthango Capital Markets-ukutyala kwimarike imfuyo ezisekwe wedatha eqokelelwe Twitter (Jordan 2010) . Yintoni bungqina kungabangela ukuba ufuna ukubona phambi kubeka imali ngaloo mali?
[ ] Nangona abanye wabathetheli kwimpilo yoluntu sisichotho e-imidiza njengoncedo olusebenzayo ngenxa yokuyeka ukutshaya, abanye lumkisa malunga nobungozi, ezifana eliphezulu-amanqanaba intshongo. Khawucinge ukuba umphandi ugqiba ukufunda uluvo loluntu ngokuphathelele e-imidiza ngokuqokelela izithuba ezinxulumene-e-imidiza Twitter nokuqhuba uhlalutyo ziimvakalelo.
[ ] NgoNovemba 2009, Twitter watshintsha umbuzo kwibhokisi tweet ukusuka "wenza ntoni?" Kwi "Kwenzeka ntoni?" (Https://blog.twitter.com/2009/whats-happening).
[ ] Kwak et al. (2010) zahlalutywa izigidi 41.7 profiles yomsebenzisi, nobudlelwane kwezentlalo 1.47 billion, 4262 Topics, kwaye izigidi 106 Tweets phakathi kuJuni 6 noJuni 31, 2009. Ngokusekelwe kolu hlahlelo bagqiba ukuba Twitter esebenza kakhulu njengolwimi entsha ukwabelana ngolwazi ngaphezu inethiwekhi yokuncokola.
[ ] "Retweets" adla ngokusetyenziswa ukulinganisa impembelelo wayaneka yempembelelo on Twitter. Ekuqaleni, abasebenzisi kwafuneka ekukhupheleni ukucola tweet abayithandayo, ilabhuli umbhali yoqobo kunye / nomqheba yakhe, kwaye ngesandla chwetheza "RT" phambi tweet ukubonisa ukuba yeyona utwithe. Emva koko, ngo-2009 Twitter wongezelela "utwithe" iqhosha. NgoJuni 2016, Twitter wenza kube lula kubasebenzisi ukuba utwithe tweets zabo (https://twitter.com/twitter/status/742749353689780224). Ngaba ucinga ukuba ezi nguquko kuchaphazela yokusebenzisa "favorites" kuphando yakho? Ngoba kutheni?
[ , , ] Michel et al. (2011) sakhiwe Nenkcubeko asakhulayo ukusuka iinzame Google ukuba ziseke iincwadi. Ukusebenzisa inguqulelo yokuqala sixa, eyapapashwa ngo-2009 kwaye equlethwe iincwadi ezingaphezu kwezigidi 5 ekhompyutheni, ababhali zahlalutywa ilizwi kusetyenziso rhoqo ukuphanda utshintsho yeelwimi kunye neentsingiselo nenkcubeko. Kungekudala Google Books Corpus waba umthombo wolwazi ethandwayo kubaphandi, nenguqulelo ye2 ledatha wakhululwa ngowe-2012.
Nangona kunjalo, Pechenick, Danforth, and Dodds (2015) walumkisa ukuba abaphandi kufuneka ukuchaza ngokupheleleyo inkqubo zeesampuli sixa phambi kokusebenzisa ukuze izigqibo ezibanzi. Umba ophambili kukuba sixa na ilayibrari-onjengawe, equlethe enye incwadi nganye. Ngenxa yoko, umntu, yombhali uyakwazi ukufaka ngokuphawulekayo amabinzana abatsha yeembekiselo Iincwadi Google. Ngaphezu koko, imibhalo yenzululwazi saveza onyuka substantive inxalenye sixa kulo 1900s. Ukongeza, ngokuthelekisa iinguqulelo ezimbini IsiNgesi iiseti Bubuxoki, Pechenick et al. bafumana ubungqina bokuba ukuhluza zaneleyo yasetyenziswa ukuvelisa inguqulelo yokuqala. Zonke iinkcukacha ezifunekayo ukuze umsebenzi ifumaneka apha: http://storage.googleapis.com/books/ngrams/books/datasetsv2.html
[ , , , ] Penney (2016) phonononga ingaba kuluntu ngokubanzi malunga NSA / prism esweni (oko kukuthi, nezityhilelo Snowden) ngoJuni 2013 esidityaniswe ukuhla esibukhali kwaye ngesiquphe yezithuthi kumanqaku Wikipedia kwizihloko ezithile iinkxalabo zabucala. Ukuba kunjalo, olu tshintsho kwindlela yokuziphatha kuya kufana kunye nempembelelo ubhobhoyi ngenxa yesifo ubunzima. Indlela Penney (2016) wambi ebizwa ngokuba waphazamiseka kuyilo ixesha uthotho kwaye enxulumene ezinkqubo kwisahluko malunga approximating experimenti ezininzi eziphuma data lokuqwalasela (Icandelo 2.4.3).
Ukukhetha amagama angundoqo isihloko, Penney luthunyelwa kuluhlu esetyenziswa US Department of Lezokhuseleko ngenxa yokulandelela nokuhlola eendaba kwezentlalo. Uluhlu zeSithili hlela amagama athile uphendlo ibe uluhlu lwemiba, ie "Health Inkxalabo," "Ukhuseleko Infrastructure," yaye "Ubugrogrisi." Iqela lofundo, Penney wasebenzisa angudoqo mane anesibhozo ezinxulumene "Ubugrogrisi" (jonga Uluhlu 8 kwisihlomelo). Koko oluqinile Wikipedia inqaku imbono lixhomekeke qho ngenyanga engqinelana Wikipedia amanqaku amane anesibhozo kwisithuba seenyanga ezingamashumi amathathu anambini, ukususela ekuqaleni ngoJanuwari 2012 ukuya ekupheleni kuka-Agasti 2014. Ukuze ingxoxo yakhe, naye wadala uthelekiso ezininzi amaqela ngokuthi kubekwe inqaku izimvo malunga neminye imiba.
Ngoku, uza ukuyiphinda nokwandisa Penney (2016) . Yonke idata ekrwada ukuba kuya kufuneka ukuba lo msebenzi luyafumaneka Wikipedia (https://dumps.wikimedia.org/other/pagecounts-raw/). Okanye ungafumana ukusuka wikipediatrend R impahla (Meissner and Team 2016) . Xa ubhala-up iimpendulo zakho, nceda uqaphele apho umthombo idata kusetyenziswa. (Qaphela: Lo msebenzi mnye kuyathethwa kwiSahluko 6)
[ ] Efrati (2016) iingxelo, ngokusekelwe kulwazi oluyimfihlo, ukuba "ukwabelana iyonke" on Facebook kwehla ngeminyaka malunga 5.5% ngaphezu konyaka ngelixa "yokwabelana yosasazo yantlandlolo" ezantsi unyaka phezu konyaka-21%. Le yokwaliwa etsolo ingakumbi nabasebenzisi Facebook abangaphantsi kweminyaka engama-30 ubudala. Ingxelo kubangelwa mayinganxaxheli izinto ezimbini. Enye ukukhula kwamanani "abahlobo" abantu on Facebook. Eyesibini kukuba umsebenzi yokwabelana iye yasiwa imiyalezo kwaye abadlali ezifana ngene. Ingxelo ngokusichazela amaqhinga ezininzi Facebook bazama ukuba kunyuswe yokwabelana, kuquka News Feed lweteksti algorithm ezenza izithuba original yabalasela ngakumbi, ngokunjalo izikhumbuzo ngamaxesha abasebenzisi izithuba original "On This Day" Kwiminyaka embalwa edlulileyo. Yintoni iziphumo, ukuba ikho, ngaba ezi ziphumo kubaphandi abafuna ukusebenzisa Facebook njengomthombo data?
[ ] Tumasjan et al. (2010) bachaza ukuba umlinganiselo tweets ukukhankanya iqela lezopolitiko ufanise masilivakalise leevoti ezizuzwe kunyulo waseJamani yepalamente 2009 (Figure 2.9). Ngamanye amazwi, kwabonakala ukuba ukuba usebenzise Twitter ukuqikelela unyulo. Ngelo xesha esi sifundo yapapashwa yayigqalwa imincili kakhulu kuba kwabonakala ukuba bacebisa ukusetyenziswa esibalulekileyo umthombo oqhelekileyo data ezinkulu.
Kunikwa iimpawu ezimbi data ezinkulu, nangona kunjalo, kufuneka ngoko nangoko sokuthandabuza le japan. AmaJamani on Twitter ngo-2009 eziliqela iqela non-ummeli, yaye sixhasa iqela elinye ukuze GamaLamLomsebenzisi kwezopolitiko rhoqo. Ngoko, kubonakala akumangalisi ukuba zonke bevinjwa kangangoko unako ukuba ucinge ukuba ngandlel 'zizakucimana phandle. Enyanisweni, iziphumo Tumasjan et al. (2010) yajika entle kakhulu ukuba kuyinyaniso. Xa iphepha yabo, Tumasjan et al. (2010) ingqalelo amaqela ezopolitiko ezintandathu: Democrats Christian (CDU), Christian Social Democrats (CSU), SPD, benkululeko (FDP), ekhohlo (Die Linke), kunye Green Party (Grüne). Nangona kunjalo, iqela lezopolitiko eyona ekhankanywe isiJamani on Twitter ngelo xesha Pirate Party (Piraten), iqela elilwa ummiselo urhulumente Internet. Xa Pirate Party kuqukiwe kuhlahlelo, Twitter ukhankanya uba kunokukhokelela eyoyikekayo kweziphumo zonyulo (Figure 2.9) (Jungherr, Jürgens, and Schoen 2012) .
Kamva, abanye abaphandi ehlabathini lonke baye basebenzisa iindlela-ezifana fancier njengoko ngokusebenzisa uhlalutyo luvo ukwahlula phakathi ezintle nezimbi ukhankanya zamaqela-ukwenzela ukuphucula amandla data Twitter ukuqikelela iintlobo ezahlukeneyo kunyulo (Gayo-Avello 2013; Jungherr 2015, Ch. 7.) . Nantsi indlela Huberty (2015) esishwankathela iziphumo zezi amalinge ukuqikelela unyulo:
"Zonke iindlela nokuqikelelwa eyaziwa ngokusekelwe eendaba ntle baye basilela xa phantsi neemfuno lokwenyaniso phambili-ejonge nolwemo- lonyulo. Oku kusilela kubonakala ngenxa iimpawu ezibalulekileyo ngamajelo zentlalo, kunokuba nobunzima oqingqiweyo okanye algorithmic. Ngokufutshane, amajelo zentlalo musa, yaye mhlawumbi engasoze, wosondeza esitalini, engenamkhethe, picture ummeli labavoti; kunye neesampuli lula eendaba zentlalo aswele data eyaneleyo ukulungisa ezi ngxaki kwizithuba zokukhawulelana nemeko. "
Funda ezinye uphando ezikhokelela Huberty (2015) kweso sigqibo, ubhale memo page omnye umviwa kwezopolitiko echaza ukuba nendlela Twitter kufuneka isetyenziselwe kufunga kunyulo.
[ ] Yintoni umahluko phakathi kwezentlalo kunye mbali? Ngokutsho Goldthorpe (1991) , umahluko ophambili phakathi kwezentlalo kunye mbali ulawulo phezu ukuqokelelwa kwedatha. Mbali banyanzelwa ukuba usebenzise izinto kanti abaziingcali kwezentlalo ungayitshintsha nokuqokelela iinkcukacha kwiinjongo ezithile. Funda Goldthorpe (1991) . Njani umahluko phakathi wenzululwazi kunye nembali enxulumene ingcamango Custommades kunye Readymades?
[ ] Ukwakha phezu kumbuzo ongasentla, Goldthorpe (1991) wazoba inani iimpendulo ezibalulekileyo, kuquka omnye Nicky Hart (1994) ukuba umngeni ukuzinikela Goldthorpe ukuba alungiselele data zenziwe. Ukucacisa umda ezinokwenzeka data zazenzelwe, Hart wachaza Worker Project Affluent, uphando elikhulu ukulinganisa ubudlelwane phakathi udidi lwentlalo kwaye ukuvota ukuba lwenziwa yi Goldthorpe nabalingane phakathi kwiminyaka yoo-1960. Njengoko umntu oyilindeleyo umphengululi ababefuna yenzelwe data phezu bafumana data, iProjekthi Worker Affluent eqokelelwe data ukuba sanelisa ukujongana ithiyori kutshanje ecetywayo ngekamva udidi lwentlalo kwixesha yokwandisa imigangatho yokuphila. Kodwa, Goldthorpe noogxa ngandlel 'walibala "ukuqokelela ulwazi malunga nokuziphatha yokuvota yabafazi. Nantsi indlela Nicky Hart (1994) lushwankathela lonke lomntwana:
". . . ngayo [na] kunzima ukuphepha kwisigqibo sokuba abasetyhini ezazishiyiwe ngoba le 'zazenzelwe wenza' dataset kugciniwe yi ukulandelelana paradigmatic leyo ngaphandle amava ababhinqileyo. Iqhutywa ngombono ithiyori izigxina kunye inyathelo preoccupations eyindoda. . . , Goldthorpe noogxa bakhe kwakhiwa iseti wazimisa ngokwamava kwasekuveleni kungenelwe neengcinga zabo theoretical endaweni wokubhenca ukuba uvavanyo engeyiyi ukwanela. "
Hart waqhubeka:
"Iziphumo reseach yeProjekthi Worker Affluent kusixelela okungakumbi amaxabiso masculinist wenzululwazi-kwinkulungwane ye kunokuba ukwazisa iinkqubo ezithile ngokwezinga, ezopolitiko kunye nobomi eziphathekayo."
Ngaba zikho ezinye imizekelo apho ukuqokelelwa kwedatha zazenzelwe na ucalu kwalomqokeleli zinkcukacha eyakhelwe kuyo? oko kufana njani neyokuqungaqunga algorithmic? Yintoni iziphumo ukuze oku xa abaphandi kufuneka basebenzise Readymades kwaye xa kufuneka usebenzise Custommades?
[ ] Kwesi sahluko, mna kuchasene data eqokelelwe abaphandi ukuba abaphandi kunye neerekhodi zolawulo yadalwa ziinkampani kunye noorhulumente. Abanye abantu bathi ezi rekhodi zolawulo "wafumana data," leyo kwahluke "data yenzelwe." Kuyinyaniso ukuba iirekhodi zolawulo zifunyanwa abaphandi, kodwa ke zenzelwe kakhulu. Umzekelo, iinkampani tech yanamhlanje bachitha imali ixesha elininzi kunye nezibonelelo ukuqokelela kongamela data zabo. Ngoko ke, ezi ngxelo zolawulo omabini afumaneka nesilungiselelwe, nto nje kuxhomekeke imbono yakho (Figure 2.10).
Nika umzekelo umthombo data apho ekubona oko zombini afumaneka nesilungiselelwe kuluncedo xa usebenzisa ukuba umthombo idatha yophando.
[ ] Kwi isincoko banyameke, Christian Sandvig kunye Eszter Hargittai (2015) uchaze iintlobo ezimbini zophando digital, apho inkqubo lwedijithali "isixhobo" okanye "into yokufunda." Umzekelo uhlobo lokuqala isifundo apho Bengtsson noogxa (2011) basebenzisa data ifowuni ephathwayo ukulandelela kwimfuduko emva kwenyikima eHaiti ngo-2010 umzekelo uhlobo lwesibini kulapho Jensen (2007) kwizifundo ukuqalwa of mobile phones lonke Kerala, India nefuthe ekusebenzeni yemarike iintlanzi. Mna andifumani oku kuluncedo kuba icacise ukuba izifundo usebenzisa imithombo yedatha yedijithal neenjongo zahlukene nokuba basebenzisa uhlobo olufanayo umthombo data. Ukwenzela ukucacisa ngakumbi kulo mdlalo, chaza izifundo ezine ukuba ndikhe ndabona: ezimbini usebenzise inkqubo yedijithali njengesixhobo amabini ukuba usebenzise inkqubo yedijithali libe yinto zokufunda. Ungasebenzisa imizekelo kwesi sahluko ukuba ufuna.