Business analytics in sport talent acquisition

DOI: 10.4018/IJBAN.290406 In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 This a icle published as an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/4.0/) which pe mi s un es ic ed use, dis ibu ion, and p oduc ion in any medium, p o ided he au ho o he o iginal wo k and o iginal publica ion sou ce a e p ope ly c edi ed. *Co esponding Au ho 1 Business Analy ics in Spo Talen Acquisi ion: Me hods, Expe iences, and Open Resea ch Oppo uni ies Rocio de la To e, Public Uni e si y o Na a e, Spain Lau a O. Cal e , Uni e si a Obe a de Ca alunya, Spain Da id Lopez-Lopez, ESADE, Spain Angel A. Juan, Uni e si a Obe a de Ca alunya, Spain h ps://o cid.o g/0000-0003-1392-1776 Sa a Ha ami, Uni e si a Obe a de Ca alunya, Spain ABSTRACT Rec ui men o young alen ed playe s is a c i ical ac i i y o mos p o essional eams in di e en spo s such as oo ball, socce , baske ball, baseball, cycling, e c. In he pas , he selec ion o he mos p omising playe s was done jus by elying on he expe s’ opinions bu wi hou sys ema ic da a suppo . Nowadays, he exis ence o la ge amoun s o da a and powe ul analy ical ools ha e aised he in e es in making in o med decisions based on da a analysis and da a-d i en me hods. Hence, mos p o essional clubs a e in eg a ing da a scien is s o suppo manage s wi h da a-in ensi e me hods and echniques ha can iden i y he bes candida es and p edic hei u u e e olu ion. This pape e iews exis ing wo k on he use o da a analy ics, a i icial in elligence, and machine lea ning me hods in alen acquisi ion. A nume ical case s udy, based on eal-li e da a, is also included o illus a e some o he po en ial applica ions o business analy ics in spo alen acquisi ion. In addi ion, esea ch ends, challenges, and open lines a e also iden i ied and discussed. KEywORdS Business Analy ics, Machine Lea ning, Spo s, Talen Acquisi ion 1. INTROdUCTION In he p esen day, inding and hi ing alen ed wo ke s has become one o he op p io i ies o many businesses. Ine icien hi ing p ac ices ha e a nega i e epe cussion on any o ganiza ion, and migh impose conside able loses, bo h in e ms o money and ime. As poin ed ou by Da enpo e al. (2010), hose companies ha a e capable o a ac ing and e aining he bes alen ed people a e among he mos compe i i e ones. Acco ding o Ha is e al. (2011), in a globalized and highly compe i i e en i onmen mos o ganiza ions should s a using da a o measu e and imp o e he con ibu ion o hei human esou ces (HR) o hei pe o mance. In he spo s sec o , Bake e al. (2017), De Bossche and De Rycke (2017) and Hanlon e al. (2014) s a e ha he e is an inc easing in e es in unde s anding he cos s and bene i s o ini ia i es In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 2 o ea ly iden i ica ion o alen ed playe s, as well as in unleashing he ac o s ha in luence a hle es’ de elopmen . These au ho s also a i m ha , while adi ional s a is ical analysis was ocused on ma ch a iables, such us goals sco ed o playe s’ posi ion on he ield, ecen ad ances in spo s analy ics a e ocused on mo e complex issues like alen acquisi ion. As poin ed ou by Ge a d (2017), he 2011 ilm ‘Moneyball’1 highligh ed he possibili ies o analy ics as a compe i i e s a egy, pa icula ly o small-ma ke eams wi h ela i ely limi ed esou ces. Following F ied and Mumcu (2016), many coaches employ da a on habi s and pe o mance indica o s o assess he po en ial o hei playe s. In ac , i is possible o use da a o: (i) e alua e playe s’ pe o mance; (ii) ank playe s (Pappala do e al., 2019); (iii) es ima e he alue o playe s in ans e ma ke s (Kim e al., 2019); (i ) loca e he bes posi ion ha a playe can occupy in he ield; o ( ) o ecas playe s’ goal sco ing pe o mance in he nex season (Apos olou and Tjo jis, 2019). As illus a ed in Figu e 1, adap ed om 21s Club2, clubs can use da a o analyze he impac o a new playe on he eam’s o e all pe o mance le el. The spo s indus y is being ans o med by da a analy ics in he ollowing dimensions: (i) a clubs le el, e.g., socce clubs like Li e pool, Ba celona3, A senal, Manches e Ci y, o Milan a e among he ones ha al eady use da a analysis o imp o e pe o mance, analyze i als, p e en inju ies, op imize he managemen o he ans e ma ke , and also he acquisi ion o new alen ; (ii) ega ding new en an s in da a managemen and analysis, new pla o ms o da a analysis and managemen appea o p o ide se ices o clubs, such as Wiscou 4 and Scispo s5; (iii) as ega ds as new en an s in da a cap u e and gene a ion, la ge companies such as In el ha e launched he c ea ion o wea able In e ne -o -Things de ices, which a e capable o cap u ing playe s’ in o ma ion in eal ime6; (i ) wi h espec o ans o ma ion o spo s managemen p o essionals, new o icial uni e si y deg ees dedica ed o spo s managemen ha e eme ged, mos o hem wi h a special emphasis on da a science; and ( ) a ans o ma ion o spo s en husias s, who ha e begun o consume complex da a, bo h o hei own digi iza ion p ocess as well as o hei leisu e and un when using online be ing applica ions. Figu e 1. Playe ecommenda ion based on da a analy ics (adap ed om 21s Club) In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 3 This pape p esen s a comp ehensi e e iew o he s a e o he a ega ding da a-d i en app oaches o alen acquisi ion in spo s. In addi ion, he pape discusses he mos common used me hods o alen acquisi ion. Some o hese me hods belong o he ields o a i icial in elligence (AI) and machine lea ning (ML), which a e also in oduced in he con ex o spo s analy ics. Finally, he pape iden i ies and discusses ends, challenges, and open esea ch lines ela ed o his esea ch a ea. The es o he pape is o ganized as ollows. Sec ion 2 b ing ou he concep o alen analy ics, specially in he spo s sec o . Sec ion 3 p o ides a sho in oduc ion o he ields o AI and ML, hus allowing he un amilia eade o ollow he ollowing sec ions. Sec ion 4 discusses di e en da a- d i en analy ic app oaches. A e wa ds, Sec ions 5 and 6 e iew esea ch pape s on da a analy ics in ec ui men and spo s alen acquisi ion, espec i ely. Sec ion 7 p o ides an o iginal case s udy based on a eal-li e da ase o Eu opean socce playe s, whe e some o he po en ial o ML me hods is illus a ed. T ends, challenges, and open esea ch lines a e discussed in Sec ion 8. Finally, Sec ion 9 summa izes he main con ibu ions o his pape . 2. TALENT ANALyTICS IN SPORTS Talen analy ics (TA) ep esen s a g oundb eaking oppo uni y o many o ganiza ions in he spo s sec o . TA is de ined by Bassi (2011) as “an e idence-based app oach o making be e decisions on he people side o he business; i consis s o an a ay o ools and echnologies, anging om simple epo ing o HR me ics all he way up o p edic i e modeling”. Du ing he las yea s, p o essional and eli e spo eams a e gi ing a conside able a en ion o implemen business-o ien ed TA in hei s a egy. Hence, o Da ids and A au´jo (2019) he main challenge is no he managemen o alen among he playe s who al eady belong o a eam, bu he acquisi ion o young and alen ed playe s o he u u e. When conside ing young people, e olu ion and o ecas ing models ha go beyond da a on he cu en pe o mance le el should be buil as well (Webb e al., 2020; Pi e , 2019; Fo d and Williams, 2017). As Williams and Reilly (2000) s a e, mul i-dimensional da a (including physical, psychical, and sociological cha ac e is ics) has o be conside ed. Likewise, Da che a (2014) main ains ha da a ob ained om social ne wo ks could be used as a p edic o o po en ial child en’s alen , while Ma in (2015) explo es he in luence o he child socio-economic s a us on i s pe o mance and u u e e olu ion. O he au ho s (Kand a´ˇc e al., 2019; Picke ing e al., 2019; Loland, 2015; Webbo n e al., 2015; Cˆo ´e, 1999) discuss abou whe he alen is an inna e skill o no , and i da a analysis should be o ien ed owa ds inna e capaci ies and genes a he han o pe o mance. Gi en he numbe o ac o s ha may a ec decisions ela ed o alen acquisi ion, au ho s such as Vaeyens e al. (2008), Ge a d (2017), and Be gkamp e al. (2019) conclude ha alen acquisi ion is a mul i-dimensional challenge, and one ha is no only based on he skills o indi idual playe s bu also on he whole eam, i.e., he cu en eam con igu a ion has o be conside ed as well. In his con ex , Ge a d (2017) p opose he use o simula ion models as mo e e ec i e ools han expe judgmen , especially in a mul i-dimensional en i onmen like he one being conside ed. Using da a analy ics, Gandelman (2009) was able o show ha ou side oppo uni ies we e highe o socce playe s wi h a supe io socioeconomic backg ound and a be e educa ion. He also ound e idence o acial dis- c imina ion in he U uguayan socce ma ke , whe e ob aining a p o essional con ac was easie o whi e playe s. Simila ly, Be i and B ook (2010) used da a analy ics o iden i y some e iciencies ega ding he e alua ion o playe s’ pe o mance in he Na ional Hockey League. Employing Bayesian analysis combined wi h Ma ko Chain Mon e Ca lo es ima ion, Rimle e al. (2010) s udied he echnical e iciency o a baske ball eam. They we e no able o ind signi ican di e ences, in echnical e iciency le els, ac oss eams playing in he same ca ego y. B yson e al. (2013) in es iga ed how he sala y o a socce playe migh depend upon his abili y o play wi h bo h ee . A e analyzing da a om he i e main Eu opean leagues, hey concluded ha he a o emen ioned playe s end o ecei e a no iceable sala y p emium. In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 4 3. ARTIFICIAL INTELLIGENCE ANd MACHINE LEARNING The popula i y o AI and ML me hods has been cons an ly inc easing du ing he las decade (Joshi, 2019). The disciplines in which hey ha e been o igina ed a e nume ous, including: Compu e Science, Ma hema ics, Elec ical Enginee ing, S a is ics, Signal P ocessing, e c. These ields ind applica ions in a wide ange o indus ies, such as image p ocessing, na u al language p ocessing, and online shopping. Figu es 2 and 3 shows he a ie y o AI- ela ed and ML- ela ed esea ch disciplines, and he impac gene a ed measu ed in e ms o he numbe o indexed publica ions in he Web o Science (WoS). Figu e 2. WoS-indexed con ibu ions in A i icial In elligence acco ding o i s esea ch a ea Figu e 3. WoS-indexed con ibu ions in Machine Lea ning acco ding o i s esea ch a ea In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 5 Acco ding o B oussa d e al. (2019), AI e e s o machines capable o pe - o ming one o mo e asks associa ed wi h he human na u e, o ins ance: lea n and p ocess human language, execu e mechanical asks ha equi e complex maneu e ing, sol ing compu e -based complex p oblems ha may in ol e la ge amoun da a in a e y sho ime lapse. O he s, like Li and Du (2017), de ine he e m as “ a ie y o human in elligen beha io s, such as pe cep ion, memo y, emo ion, judgmen , o easoning, ha can be ealized a i icially by a machine, a sys em, o a ne wo k”. Mos cu en applica ions a e ocused on he ield o neu al ne wo ks. Fo ins ance, deep neu al ne wo ks a e used o speech ecogni ion (T e e , 1997), image classi ica ion (Deepa and De i, 2011), o p edic ion o wo ds in a ex (Ba aglia e al., 2016). ML as subse o AI e e s o a compu e p og am ha can lea n how o p oduce a ce ain beha io , which was no explici ly p og ammed in i (Ko sian is e al., 2007). Indeed, i can be capable o showing beha io s om which he p og amme may be comple ely unawa e o . As in human beha io , many aspec s o lea ning and in elligence in AI a e closely ela ed o he ep esen a ion o unce ain y. The e o e, p obabilis ic app oaches a e undamen al (Ghah amani, 2015). P obabilis ic me hods y o assign an unce ain y measu e o he un- known a iables, as well as a ce ain p obabili y o known a iables. Hence, he goal is o ind he unknown alues using p obabilis ic models. These models a e classi ied in o wo main ypes, so called gene a i e and disc imina i e: disc imina i e models y o o ecas he changes in he ou pu jus conside ing he changes occu ed in he inpu , while gene a i e models a e he ones in which he changes in he ou pu can be explained as a consequence o changes in he inpu as well as changes in he s a e (Joshi, 2019). Cu en esea ch ega ding he on ie o p obabilis ic ML app oaches (bo h disc imina i e as well as gene a i e) is mainly ocused on: (i) p obabilis ic p og amming (as a gene al amewo k o exp essing p obabilis ic models as compu e p og ams); (ii) Bayesian op imiza ion ( o globally op imizing unknown unc ions); (iii) hie a chical modeling o lea ning many ela ed models; and (i ) p obabilis ic da a comp ession (Ghah amani, 2015). ML app oaches can be di ided in o supe ised and unsupe ised lea ning. The o me a e in ol ed in many applica ions and deal wi h p oblems ela ed o lea ning wi h guidance. In o he wo ds, he aining da a in supe ised lea ning me hods needs labeled samples. Thus, o ins ance, samples wi h class labels a e equi ed in a classi ica ion p oblem. Hence, he ma hema ical model lea ns i s pa ame e s om labeled samples wi h he main goal o making p edic ions on samples ha he model has no seen be o e. Then, he classi ie is used o assigning class labels o he es ing ins ances in which he alues o he p edic o ea u es a e known, bu he alue o he class label emains unknown. Since he supe ised classi ica ion is one o he asks equen ly de eloped by ‘in elligen sys ems’, i seems logical ha a g ea numbe o echniques a e based on AI and s a is ics. Meanwhile, unsupe ised lea ning deals wi h p oblems ha in ol e da a wi hou labels. In his case, he machine ecei es inpu s bu ob ains nei he ou pu s no ewa ds om i s en i onmen (Ghah amani, 2003). Unsupe ised app oaches y o ind ends and some kind o s uc u e in he aining da a. Tha is, hese app oaches y o unde s and he o igin o he da a i sel and o build ep esen a ions o he inpu s ha can be used o decision-making, e icien ly communica ing he inpu s o ano he machine o p edic ing u u e inpu s. Clus e ing is a ypical example o unsupe ised lea ning. In unsupe ised lea ning, he majo i y o he wo k can be conside ed as a lea ning p ocess o a p obabilis ic model. When he scena io is no able o gi e he machine any supe ision o ewa d, he machine can design a model ha ep esen s he p obabili y dis ibu ion o a new inpu . This is achie ed by jus conside ing a p e ious use ul inpu (e.g., s ock p ices o wea he condi ions). P obabilis ic models ha can be used in unsupe ised lea ning a e, among o he s: ac o analysis, independen componen s analysis, p incipal componen s analysis, o Gaussians models. The e a e si ua ions whe e he supe ised me hods a e no he bes op ion. The i s and mos impo an is he high cos o labeling. Mo eo e , ha ing all he aining da a ully labeled can be p ac ically impossible. In hese cases, i is common o s a wi h supe ised me hods –using a small In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 6 se o labeled da a–, and hen imp o e he model in an unsupe ised way–i.e., using a la ge se o unlabeled da a. AI and ML echniques a e p esen in almos all sec o s, including spo s. In ac , mos a iables ha can be quan i ied can also be p edic ed using AI and ML. The spo sec o is ull o quan i iable elemen s, which makes i ideal o he use o hese echniques. Fo example, ec ui men o playe s is one o he a eas in spo s whe e AI and ML a e inc easingly employed (Cha an, 2019; He old e al., 2019; Musa e al., 2019; C´wiklinski e al., 2021). 4. TyPES OF ANALySIS IN TALENT ANALyTICS Da a analy ic app oaches, which a e ypically based on ML and s a is ics me hods, can b ing insigh s ha a e c i ical o imp o ing ope a ional and business ou comes o many o ganiza ions. These app oaches play a ole as powe ul ools in he seeking and hi ing young alen . Talen analy ics is a sys ema ic p ocess ha applies s a is ics, echnology, and expe ise o la ge se s o people da a o disco e he meaning ul pa e ns ha allow o sup- po ing decision-making in ec ui men . Th ee common ypes o analy ics –desc ip i e, p edic i e, and p esc ip i e– a e used in TA, people and human esou ce analy ics amewo ks o measu e e iciency, e ec i eness, quali y o ec ui men , and impac (Necula and S imbei, 2019). F om desc ip i e o p esc ip i e analysis, no only he model inc eases in he complexi y o he da a being used, bu also he analysis p og essi ely ge s mo e sophis ica ed. Each o he a o emen ioned ypes a e desc ibed nex : Desc ip i e Analy ics: epo ing / isualiza ion is he i s s ep o ca ying ou s a is ical analyses, and i is used o desc ibe he basic ea u es o he da a ia applying o he collec ed da a h ough he s uc u ed ques ionnai e (Ma ybe h e al., 2019); i plays a ele an ole in p o iding a iew in o ac i i y –such as equisi ion olume, alen pool size, sou ce o hi es, e c.–, as well as o e eal he le els o ac i i y and e iciency in each candida e gene a ion. P edic i e Analy ics: uses da a o ind pa e ns and employs hem o p edic he u u e; i pe mi s o iden i y s a is ical ela ionships be ween a se o ac i i ies and he expec ed ou comes, ha will help o: (i) o ecas wha will happen in he u u e o o explain he ob ained ou come (like a candida e’s likely cul u al i , le el o pe o mance, and e en ion); o (ii) no ice po en ial alen sho ages o skills gaps, and ma ke a ailabili y (wo k o ce planning); also, p edic i e analy ics is mainly ela ed o selec ion o ejec ion o candida e, accep ance o p o ided o e by selec ed candida es and oo cause analysis o o e decline (S i as a a e al., 2015); p edic i e analysis inds answe s o ques ions such as ‘wha will he u u e look like?’, ‘wha ac ics mos in luence business ou comes?’, e c.; he esul o his p edic i e analysis help sho en he en i e ec ui ing p ocess while imp o ing he hi ing p ocess. P esc ip i e Analy ics: goes a s ep u he in he u u e and a emp s o p o ide and sugges be e decisions using da a echniques such as decision modeling, ML, heu is ic, simula ion, neu al ne wo ks; hese decisions a e based on he esul s p o ided by p edic i e analysis; i ies o e alua e he e ec and impac o he p o ided decisions in o de o modi y hem be o e implemen a ion; i usually esul s in ules and ecommenda ion o nex s eps (A a an and A a an, 2018); o example alen acquisi ion eam ecei es he hi e o no hi e sugges ions o s a egy ecommenda ion om p edic i e analysis. A common da a analy ics amewo k used in TA is shown in Figu e 4. The mos common da a-d i en me hodologies used in TA may classi ied in In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 7 eg ession, classi ica ion, clus e ing, associa ion ule mining, and anomaly de ec ion. The eade in e es ed in a mo e comple e in oduc ion o TA is e e ed o Da enpo e al. (2010), which illus a es uses and desc ibes he undamen als o build a capaci y in his domain, i.e.: access o high-quali y da a, en e p ise o ien a ion, analy ical leade ship, and s a egic a ge s. In his con ex , Nocke and Sena (2019) discusses he ad an ages and cos s induced (in e ms o da a go e nance and e hics) by using TA wi hin an o ganiza ion. The au ho s p esen a numbe o case s udies o analyze he posi i e e ec o TA usage in o ganiza ional decision-making p ocesses and de e mine he key channels h ough which he TA adop ion imp o e HR managemen and, subsequen ly, he whole o ganiza ion unc ion. 5. dATA ANALyTICS IN RECRUITMENT A mo e da a-d i en cul u e is becoming inc easingly popula among companies and go e nmen s. HR cons i u es an example o a business’ depa men ha has d ama ically changed du ing he las decades due o he use o da a analy ics me hodologies and echnologies. Indeed, companies a e inc easingly adop ing sophis ica ed me hods o s udy employee’s da a in o de o imp o e he decision-making p ocess, so hey can s eng h hei compe i i e ad an age (Da enpo e al., 2010). Acco ding o Rana e al. (2019), TA shows he po en ial wi hin he decisions ega ding hi ing, aining, imp o ing p oduc i i y, and e aining alen , all o hem wi h he main pu pose o make a company mo e compe i i e. Ga ne , Inc.7, poin s ou ha he olume o da a and me ics a ailable o HR has inc eased exponen ially, while 70% o companies expec o in- c ease he esou ces hey dedica e o TA in he coming yea s. E en so, only 21% o he HR leade s belie e ha hei o ganiza ions a e e ec i e a using alen da a o in o m business decisions. Dey and De (2015) poin s ou i e key a eas whe e p edic i e analy ics can c ea e alue in HR: (i) employee p o iling and segmen a ion, employee a i ion, and loyal y analysis; (ii) o e- cas ing o HR capaci y and ec ui men needs; (iii) app op ia e ec ui men p o ile selec ion; (i ) employee sen imen analysis; and ( ) employee aud isk managemen . Figu e Figu e 4. Da a analy ics amewo k using in Talen Analy ics In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 8 5 lis s he main TA applica ions, me hodologies, and echnologies. I is based on Kau and Fink (2017), which o e s a e iew o key app oaches, compe encies and ools, building on 22 in e iews wi h academics, consul an s and p ac i ione s a 16 co po a ions, as well as on o he TA expe s. 5.1. Rec ui men and Talen Acquisi ion Acco ding o Wikipedia, ec ui men may be de ined as “ he p ocess o a ac ing, sho lis ing, selec ing, and appoin ing sui able candida es o jobs wi hin an o ganiza ion, and is a key unc ion o human esou ce managemen .” Ano he in e es ing de ini ion is p o ided by B eaugh (2008), which de ines ex e nal ec ui men as “an employe ’s ac ions ha a e in ended o: (i) b ing a job opening o he a en ion o po en ial job candida es who do no cu en ly wo k o he o ganiza ion; (ii) in luence whe he hese indi iduals apply o he opening; (iii) a ec whe he hey main ain in e es in he posi ion un il a job o e is ex ended; and (i ) in luence whe he a job o e is accep ed”. Rec ui men plays an essen ial ole in de e mining he e ec i eness o o ganiza ions, and i is composed o se e al sub-p ocesses: (i) job analysis, which consis s in documen ing he knowledge, skills, abili ies, and o he cha ac e is ics (KSAOs) equi ed o a job; (ii) sou cing, which is he p ocess o a ac ing o iden i ying candida es; and inally (iii) sc eening and selec ion. O ganiza ions apply ec ui men s a egies o iden i y hi ing acancy, es ablish imelines, and de ine goals h oughou he ec ui men p ocess. Each o ganiza ion designs i s own s a egies o ec ui men , bu he e a e equen app oaches such as using social ne wo ks o ex e nal ec ui men ad e isemen o employing s anda d psychological es s, g oup discussion and a numbe o in e iews o assess a a ie y o KSAOs. Thus, he ec ui men p ocess is complex and equi es big amoun s o e o and in es men . Bha acha yya (2015) de ines alen acquisi ion as a s a egic app oach aiming o iden i y, a ac , and b ing onboa d op alen o mee dynamic business needs. Acco ding o his au ho , ec ui ing is mo e ac ical and ocuses mos ly on immedia e hi ing needs, i.e., a p ocess o illing he open posi ions. Figu e 6 allows us o check he g owing popula i y o his esea ch ield. Mo e speci ically, i shows he e olu ion, om 2000 o 2019, in he numbe o wo ks epo ed by he Web o Science when sea ching o : (i) “Talen acquisi ion” o “Talen ec ui men ” (which some imes a e used as synonyms in he li e a u e) in “Social Science” (in o ange); and (ii) he same bu in “spo s” (in blue). Clea ly, he e has been a posi i e and sus ained end du ing he las 10 o 15 yea s. 5.2. S udies on da a-d i en Rec ui men He e, we desc ibe a ew ecen and ep esen a i e wo ks on ec ui men using da a-in ensi e me hodologies and echniques. Fo ins ance, Mohapa a and Sahu (2017) p esen s a case s udy o highligh he misconcep ions in capaci y planning me hods (i.e. hi ing p ocess) and in he de ec ion o bo le- necks in he hi ing pipeline by using ec ui men unnel echnique and some me ics o check e iciency o he a o emen ioned hi ing p ocess. Mo eo e , he au ho s depic app op ia e sou ces o hi ing based on he pe o mance o candida es hi ed om hose sou ces. They iden i y success ul p o iles in he company h ough compu ing co ela ions be ween selec ion pa ame e s and pe o mance sco es. Finally, hey p o ide a ec ui men s a egy and u u e oad map o he case s udy. Kau and Fink (2017) in es iga es wha is e- qui ed o se up and un an e ec i e TA unc ion and he s uc u es, sys em and skills ha enable i . The au ho s discuss he answe o hese ques ions h ough he collec ion and analysis o da a om 22 in e iews wi h academics, consul an s, and di e en indus ies. The analysis shows ha da a in as uc u e and epo ing, ad anced analy ics, and o ganiza ional esea ch a e h ee componen s o a ma u e TA unc ion. Aza e al. (2013) p o ides a decision-making ool o help manage s du ing he ec ui men p ocess. Au ho s claim ha he ool, h ough using da a mining echniques, is able o disco e pa e ns In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 9 Figu e 5. Scheme o alen analy ics. Sou ce: based on Kau and Fink (2017) Figu e 6. E olu ion o he numbe o ela ed wo ks om 2000. Da a sou ce: Web o science. Dashed lines ep esen he endencies calcula ed as mo ing a e ages. In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 16 REFERENCES Alama , B. C. (2013). Spo s analy ics: A guide o coaches, manage s, and o he decision make s. Columbia Uni e si y P ess. Apos olou, K., & Tjo jis, C. (2019). Spo s analy ics algo i hms o pe o mance p edic ion. In 10 h In e na ional Con e ence on In o ma ion, In elligence, Sys ems and Applica ions (IISA). IEEE. A a an, M., & A a an, S. (2018). Oppo uni ies and challenges o implemen ing p edic i e analy ics o compe i i e ad an age. In e na ional Jou nal o Business In elligence Resea ch, 9. Aza , A., Rajaeian, A., Seb , M. V., & Ahmadi, P. (2013). A model o pe sonnel selec ion wi h a da a mining app oach: A case s udy in a comme cial bank. SA Jou nal o Human Resou ce Managemen , 11(1), 1–10. doi:10.4102/sajh m. 11i1.449 Babbi , D. G. (2019). A de elopmen model o guide he ec ui ing o emale sho pu e s a he NCAA Di ision I Championship le el. Jou nal o Spo s Analy ics, 5(3), 181–190. doi:10.3233/JSA-180275 Baboo a, R., & Kau , H. (2019). P edic i e analysis and modelling oo ball esul s using machine lea ning app oach o English p emie league. In e na ional Jou nal o Fo ecas ing, 35(2), 741–755. doi:10.1016/j. ij o ecas .2018.01.003 Bake , J., Cobley, S., Scho e , J., & Wa ie, N. (2017). Talen iden i ica ion and de elopmen in spo : An in oduc ion. In Rou ledge Handbook o Talen Iden i ica ion and De elopmen in Spo (pp. 1–8). Rou ledge. doi:10.4324/9781315668017-1 Bake , R. E., & Kwa le , T. (2015). Spo analy ics: Using open sou ce logis ic eg ession so wa e o classi y upcoming play ype in he NFL. Jou nal o Applied Spo Managemen , 7. Ba on, D., Ball, G., Robins, M., & Sunde land, C. (2020). Iden i ying playing alen in p o essional oo ball using a i icial neu al ne wo ks. Jou nal o Spo s Sciences, 38(11-12), 1–10. doi:10.1080/02640414.2019.17 08036 PMID:31941425 Bassi, L. (2011). Raging deba es in HR analy ics. People and S a egy, 34, 14. Ba aglia, P., Pascanu, R., Lai, M., & Rezende, D. J. (2016). In e ac ion ne wo ks o lea ning abou objec s, ela ions and physics. Ad ances in Neu al In o ma ion P ocessing Sys ems, 4502–4510. Beckwi h, G., Callahan, T., Ca lson, B., Fond en, T., Ha is, R., Hoege, J., Ma in, T., Menna, C., Summe , E., & Sche e , W. (2019). Sys ems analysis o uni e si y o i ginia oo ball ec ui ing and pe o mance. In 2019 Sys ems and In o ma ion Enginee ing Design Symposium (SIEDS). IEEE. doi:10.1109/SIEDS.2019.8735611 Be gkamp, T. L., Niessen, A. S. M., den Ha igh, R. J., F encken, W. G., & Meije , R. R. (2019). Me hodological issues in socce alen iden i ica ion esea ch. Spo s Medicine (Auckland, N.Z.), 49(9), 1–19. doi:10.1007/ s40279-019-01113-w PMID:31161402 Be i, D. J., & B ook, S. L. (2010). On he e alua ion o he “mos impo an ” posi ion in p o essional spo s. Jou nal o Spo s Economics, 11(2), 157–171. doi:10.1177/1527002510363097 Bhanda i, I., Cole , E., Pa ke , J., Pines, Z., P a ap, R., & Ramanujam, K. (1997). Ad anced scou : Da a mining and knowledge disco e y in NBA da a. Da a Mining and Knowledge Disco e y, 1(1), 121–125. doi:10.1023/A:1009782106822 Bha acha yya, D. K. (2015). The magne ic o ganiza ion: A ac ing and e aining he bes alen . SAGE Publica ions India. B eaugh, J. A. (2008). Employee ec ui men : Cu en knowledge and impo - an a eas o u u e esea ch. Human Resou ce Managemen Re iew, 18(3), 103–118. doi:10.1016/j.h m .2008.07.003 B oussa d, M., Diakopoulos, N., Guzman, A. L., Abebe, R., Dupagne, M., & Chuan, C. H. (2019). A i icial in elligence and jou nalism. Jou nalism & Mass Communica ion Qua e ly, 96(3), 673–695. doi:10.1177/1077699019859901 In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 17 B own, M. R., Delau, S., & Desgo ces, F. D. (2010). E o egula ion in owing aces depends on pe o mance le el and exe cise mode. Jou nal o Science and Medicine in Spo , 13(6), 613–617. doi:10.1016/j. jsams.2010.01.002 PMID:20227342 B yson, A., F ick, B., & Simmons, R. (2013). The e u ns o sca ce alen : Foo edness and playe emune a ion in eu opean socce . Jou nal o Spo s Economics, 14(6), 606–628. doi:10.1177/1527002511435118 Cha an, A. (2019). Rec ui men o Sui able Foo ball Playe by using Machine Lea ning Techniques (Ph.D. hesis). Na ional College o I eland. Cokins, G., & Sch ade , D. (2017). The spo s analy ics explosion. OR/MS Today, 44, 28–33. C^o ’e, J. (1999). The in luence o he amily in he de elopmen o alen in spo . The Spo Psychologis , 13(4), 395–417. doi:10.1123/ sp.13.4.395 C´wiklinski, B., Gie-lczyk, A., & Cho a’s, M. (2021). Who will sco e? a machine lea ning app oach o suppo ing oo ball eam building and ans e s. En opy (Basel, Swi ze land), 23(1), 90. doi:10.3390/e23010090 PMID:33435241 Da che a, P. (2014). Iden i ying spo s alen s by social media mining as a ma ke ing ins umen . In Annual SRII global con e ence. IEEE. doi:10.1109/SRII.2014.38 Da enpo , T. H., Ha is, J., & Shapi o, J. (2010). Compe ing on alen analy ics. Ha a d Business Re iew, 88, 52–58. PMID:20929194 Da ids, K., & A au’jo, D. (2019). Inna e alen in spo : Bewa e o an o ganismic asymme y–commen on Bake & Wa ie. In Cu en Issues in Spo Science. CISS. De Bossche , V., & De Rycke, J. (2017). Talen de elopmen p og ammes: A e ospec i e analysis o he age and suppo se ices o alen ed a hle es in 15 na ions. Eu opean Spo Managemen Qua e ly, 17(5), 590–609. doi:10.1080/16184742.2017.1324503 Deepa, S., & De i, B. A. (2011). A su ey on a i icial in elligence app oaches o medical image classi ica ion. Indian Jou nal o Science and Technology, 4(11), 1583–1595. doi:10.17485/ijs /2011/ 4i11.35 Dey, T., & De, P. (2015). P edic i e analy ics in HR: A p ime . A Whi e Pape , Ta a Consul ancy Se ices. A ailable: h p://www. cs.com/Si eCollec ionDocumen s/Whi e-Pape s/P edic i e- Analy ics-HR-0115-1.pd D onyk-T ospe , T., & S i zel, B. (2017). Lock-in and eam e ec s: Rec ui ing and success in college oo ball a hle ics. Jou nal o Spo s Economics, 18(4), 376–387. doi:10.1177/1527002515577885 Edelmann-Nusse , J., Hohmann, A., & Hennebe g, B. (2002). Modeling and p edic ion o compe i i e pe o mance in swimming upon neu al ne wo ks. Eu opean Jou nal o Spo Science, 2(2), 1–10. doi:10.1080/17461390200072201 Fo d, P. R., & Williams, A. M. (2017). Spo ac i i y in childhood: Ea ly specializa ion and di e si ica ion. In Rou ledge Handbook o Talen Iden i i- ca ion and De elopmen in Spo (pp. 116–132). Rou ledge. doi:10.4324/9781315668017-9 F ied, G., & Mumcu, C. (2016). Spo analy ics: A da a-d i en app oach o spo business and managemen . Taylo & F ancis. doi:10.4324/9781315619088 Gandelman, N. (2009). Selec ion biases in spo s ma ke s. Jou nal o Spo s Economics, 10(5), 502–521. doi:10.1177/1527002509332237 Ga iao, L. O., San ’Anna, A. P., Al es Lima, G. B., & de Almada Ga cia, P. A. (2020). E alua ion o socce playe s unde he moneyball concep . Jou nal o Spo s Sciences, 38(11-12), 1221–1247. doi:10.1080/026404 14.2019.1702280 PMID:31876264 Ge a d, B. (2017). The ole o analy ics in assessing playing alen . In Rou ledge Handbook o Talen Iden i ica ion and De elopmen in Spo . Rou ledge. doi:10.4324/9781315668017-30 Ghah amani, Z. (2003). Unsupe ised lea ning. In Summe School on Ma- chine Lea ning (pp. 72–112). Sp inge . Ghah amani, Z. (2015). P obabilis ic machine lea ning and a i icial in elli- gence. Na u e, 521(7553), 452–459. doi:10.1038/na u e14541 PMID:26017444 In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 18 Giu ida, M. (2014). Unleashing he powe o alen analy ics in ede al go e nmen . Public Managemen , 43, 7. Hanlon, C., Mo is, T., & Nabbs, S. (2014). P og am p o ide s’ pe spec i e: Rec ui men and e en ion s a egies o women in physical ac i i y p o- g ams. Spo Managemen Re iew, 17(2), 133–144. doi:10.1016/j. sm .2013.04.001 Ha is, J. G., C aig, E., & Ligh , D. A. (2011). Talen and analy ics: New ap- p oaches, highe ROI. The Jou nal o Business S a egy, 32(6), 4–13. doi:10.1108/02756661111180087 He old, M., Goes, F., Nopp, S., Baue , P., Thompson, C., & Meye , T. (2019). Machine lea ning in men’s p o essional oo ball: Cu en applica ions and u u e di ec ions o imp o ing a acking play. In e na ional Jou nal o Spo s Science & Coaching, 14(6), 798–817. doi:10.1177/1747954119879350 Joshi, A. V. (2019). Machine Lea ning and A i icial In elligence. Sp inge . Kau , J., & Fink, A. A. (2017). T ends and p ac ices in alen analy ics. Socie y o Human Resou ce Managemen (SHRM)-Socie y o Indus ial-O ganiza ional Psychology (SIOP) Science o HR Whi e Pape Se ies 20. Kim, Y., Bui, K. H. N., & Jung, J. J. (2019). Da a-d i en explo a o y app oach on playe alua ion in oo ball ans e ma ke . Concu ency and Compu a ion, 33(3), e5353. doi:10.1002/cpe.5353 Kosele , K., & S ephan, M. (2017). Machine lea ning applica ions in baseball: A sys ema ic li e a u e e iew. Applied A i icial In elligence, 31(9-10), 745–763. doi:10.1080/08839514.2018.1442991 Ko sian is, S.B., Zaha akis, I., & Pin elas, P. (2007). Supe ised machine lea ning: A e iew o classi ica ion echniques. Eme ging A i icial In elligence Applica ions in Compu e Enginee ing, 160, 3–24. Li, D., & Du, Y. (2017). A i icial in elligence wi h unce ain y. CRC P ess. doi:10.1201/9781315366951 Loland, S. (2015). Agains gene ic es s o a hle ic alen : The p imacy o he pheno ype. Spo s Medicine (Auckland, N.Z.), 45(9), 1229–1233. doi:10.1007/s40279-015-0352-5 PMID:26121951 Mankin, J.A., Ri as, J., & Jewell, J.J. (2019). The e ec i eness o college oo - ball ec ui ing a ings in p edic ing eam success: A longi udinal s udy. Resea ch in Business and Economics Jou nal, 14, 1–19. Ma ybe h, J., B ackz, J., Hadson, W., & Zlong, M. (2019). Talen acquisi ion and alen engagemen p ac ices signi ican impac o e employee sa is ac ion. In e na ional Resea ch Jou nal o Managemen , IT and Social Sciences, 6, 244–252. Ma in, L. (2015). Is socioeconomic s a us a con ibu ing ac o o ennis playe s’ success. Jou nal o Medicine and Science in Tennis, 20, 116–121. Mohapa a, M., & Sahu, P. (2017). Op imizing he ec ui men unnel in an ITES company: An analy ics app oach. P ocedia Compu e Science, 122, 706–714. doi:10.1016/j.p ocs.2017.11.427 Musa, R. M., Taha, Z., Majeed, A. P. A., & Abdullah, M. R. (2019). Machine lea ning in spo s: iden i ying po en ial a che s. Sp inge . doi:10.1007/978-981-13-2592-2 Nassib, S. H., Mkaoue , B., Riahi, S. H., Wali, S. M., & Nassib, S. (2020). P e- dic ion o gymnas ics physical p o ile h ough an in e na ional p og am e alua ion in women a is ic gymnas ics. Jou nal o S eng h and Condi ioning Resea ch, 34(2), 577–586. doi:10.1519/JSC.0000000000001902 PMID:31386634 Necula, S. C., & S imbei, C. (2019). People analy ics o seman ic web human esou ce ’esum’es o sus ainable alen acquisi ion. Sus ainabili y, 11(13), 3520. doi:10.3390/su11133520 Nocke , M., & Sena, V. (2019). Big da a and human esou ces managemen : The ise o alen analy ics. Social Sciences, 8(10), 273. doi:10.3390/socsci8100273 O oghi, B., Zeleznikow, J., MacMahon, C., & Dwye , D. (2013). Suppo ing a hle e selec ion and s a egic planning in ack cycling omnium: A s a is ical and machine lea ning app oach. In o ma ion Sciences, 233, 200–213. doi:10.1016/j.ins.2012.12.050 Pappala do, L., Cin ia, P., Fe agina, P., Massucco, E., Ped eschi, D., & Gianno i, F. (2019). Playe ank: Da a- d i en pe o mance e alua ion and playe anking in socce ia a machine lea ning app oach. ACM T ansac ions on In elligen Sys ems and Technology, 10(5), 10. doi:10.1145/3343172 In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 19 Peng, K., Cooke, J., C ocke , A., Shin, D., Fos e , A., Rue, J., Williams, R., Valei as, J., Sche e , W., & Tu le, C. (2018). P edic i e analy ics o uni e si y o i ginia oo ball ec ui ing. In Sys ems and In o ma ion Enginee ing Design Symposium (SIEDS). IEEE. doi:10.1109/SIEDS.2018.8374745 Picke ing, C., Kiely, J., G gic, J., Lucia, A., & Del Coso, J. (2019). Can gene ic es ing iden i y alen o spo ? Genes, 10(12), 972. doi:10.3390/genes10120972 PMID:31779250 Pi e , N. D. (2019). Da a analy ics in oo ball: Posi ional da a collec ion, modeling, and analysis. Jou nal o Spo Managemen , 33(6), 574–574. doi:10.1123/jsm.2019-0308 Rami ez, M. C., Vilo ia, A., Mun˜oz, A. P., & Posso, H. (2017). Applica ion o mul iple linea eg ession models in he iden i ica ion o ac o s a ec ing he esul s o he Chelsea oo ball eam. In e na ional Jou nal o Con ol Theo y and Applica ions, 10, 7–13. Rana, G., Sha ma, R., & Goel, A. K. (2019). Un a eling he powe o alen analy ics: Implica ions o enhancing business pe o mance. In Business Go e nance and Socie y (pp. 29–41). Sp inge . doi:10.1007/978-3-319- 94613-9_3 Rimle , M. S., Song, S., & Yi, D. T. (2010). Es ima ing p oduc ion e iciency in men’s ncaa college baske ball: A bayesian app oach. Jou nal o Spo s Economics, 11(3), 287–315. doi:10.1177/1527002509337803 Si a am, N., & Rama , K. (2010). Applicabili y o clus e ing and classi ica ion algo i hms o ec ui men da a mining. In e na ional Jou nal o Compu e s and Applica ions, 4(5), 23–28. doi:10.5120/823-1165 S i as a a, R., Palshika , G., & Pawa , S. (2015). Analy ics o imp o ing alen acquisi ion p ocesses. P oceedings o 4 h in e na ional con e ence on ad anced da a analysis, business analy ics and in elligence (ICADABAI 2015). T e e , D. (1997). The pa e n ecogni ion basis o a i icial in elligence. IEEE P ess. Vaeyens, R., Lenoi , M., Williams, A. M., & Philippae s, R. M. (2008). Talen iden i ica ion and de elopmen p og ammes in spo . Spo s Medicine (Auckland, N.Z.), 38(9), 703–714. doi:10.2165/00007256-200838090- 00001 PMID:18712939 Van Eck, N. J., & Wal man, L. (2010). So wa e su ey: VOS iewe , a compu e p og am o bibliome ic mapping. Scien ome ics, 84(2), 523–538. doi:10.1007/s11192-009-0146-3 PMID:20585380 Wal e , L., Ci e a, A., Knowles, K., Lowen, M., Oldenbu g, C., Shahin, H., Sche e , W., & Tu le, C. (2017). Implemen a ion o a ec ui isualiza ion ool o UVA oo ball. In Sys ems and In o ma ion Enginee ing Design Symposium (SIEDS). IEEE. doi:10.1109/SIEDS.2017.7937710 Webb, T., Dicks, M., B own, D. J., & O’Go man, J. (2020). An explo a ion o young p o essional oo ball playe s’ pe cep ions o he alen de elopmen p ocess in England. Spo Managemen Re iew, 23(3), 536–547. doi:10.1016/j.sm .2019.04.007 Webbo n, N., Williams, A., McNamee, M., Boucha d, C., Pi siladis, Y., Ahme o , I., Ashley, E., By ne, N., Campo esi, S., Collins, M., Dijks a, P., Eynon, N., Fuku, N., Ga on, F. C., Hoppe, N., Holm, S., Kaye, J., Klissou as, V., Lucia, A., & Wang, G. e al. (2015). Di ec - o-consume gene ic es ing o p edic ing spo s pe o mance and alen iden i ica ion: Consensus s a emen . B i ish Jou nal o Spo s Medicine, 49(23), 1486–1491. doi:10.1136/bjspo s-2015-095343 PMID:26582191 Williams, A. M., & Reilly, T. (2000). Talen iden i ica ion and de elopmen in socce . Jou nal o Spo s Sciences, 18(9), 657–667. doi:10.1080/02640410050120041 PMID:11043892 In e na ional Jou nal o Business Analy ics Volume 9 • Issue 1 20 Rocio de la To e is an Assis an P o esso in he Depa men o Business Managemen a he Public Uni e si y o Na a e (Spain). She is also a esea che om INARBE-Ins i u e o Ad anced Resea ch in Business and Economics. She holds a Ph.D. and a Bachelo ’s Deg ee in Indus ial Enginee ing om he Uni e si a Poli ecnica de Ca alunya. He majo esea ch a eas a e ma hema ical p og amming o s a egic planning decisions in knowledge-in ensi e o ganiza ions (KIOs) and supply chain design. Lau a O. Cal e is a Lec u e o S a is ics in he Compu e Science Dep . a he Uni e si a Obe a de Ca alunya (UOC) and Lec u e o Ma hema ics & P ojec Managemen a he Escola Uni e si à ia Salesiana de Sa ià (EUSS). She holds a M.S. in Applied S a is ics and Ope a ions Resea ch comple ed a Uni e si a Poli ècnica de Ca alunya (UPC) & Uni e si a de Ba celona (UB) and a Ph.D. in Ne wo k and In o ma ion Technologies comple ed a he UOC. She is a membe o he ICSO@IN3 esea ch g oup. He main lines o esea ch a e: - Design o op imiza ion algo i hms elying on he use o me aheu is ics, machine lea ning and/o simula ion applied o sus ainable logis ics & compu ing - Applied s a is ics & economics: applica ions in heal h, disas e managemen , & lea ning. Da id Lopez-Lopez is an academic collabo a o a ESADE. He holds a PhD in digi al ans o ma ion, and a join MBA om ESADE (Spain) and he Uni e si y o Duke (USA). He is also managing pa ne in Fhios, ha employs mo e han 180 consul an s. Angel A. Juan is a Full P o esso o Ope a ions Resea ch & Indus ial Enginee ing in he Compu e Science Dep . a he Uni e si a Obe a de Ca alunya (Ba celona, Spain). He is also he Di ec o o he ICSO esea ch g oup a he In e ne In e disciplina y Ins i u e and Lec u e a he Eunce Business School. D . Juan holds a Ph.D. in Indus ial Enginee ing and an M.Sc. in Ma hema ics. He comple ed a p edoc o al in e nship a Ha a d Uni e si y and pos doc o al in e nships a Massachuse s Ins i u e o Technology and Geo gia Ins i u e o Technology. His main esea ch in e es s include applica ions o simula ion, me aheu is ics, and machine lea ning me hods in di e en ields, including: logis ics & anspo a ion, inance, and sma ci ies. He has published o e 110 a icles in JCR-indexed jou nals and o e 250 pape s indexed in Scopus. ENdNOTES 1 h ps://en.wikipedia.o g/wiki/Moneyball_( ilm) 2 www.21s club.com 3 h ps://ba cainno a ionhub.com/es/e en /ba ca-spo s-analy ics-summi -2019/ 4 h ps://wyscou .com 5 h ps://www.scispo s.com 6 www.in el.co.uk/con en /www/uk/en/i -managemen /cloud-analy ic-hub/da a-powe ed- oo ball.h ml 7 h ps://www.ga ne .com/en/human- esou ces/insigh s/ alen -analy ics