'50, will be treated as malpractice. i : E 3 5 : = 2 : i 2 a 7 F : g § : A 2 : é 2 2 USN CCS Soneiile | 18CS72 Seventh Semester B.E. Degree Examination, Feb./Mar.2022 Big Data Analytics ‘Time: 3 hrs. Max. Marks: 100 Note: Answer any FIVE full questions, choosing ONE full question from each module. Modu Discuss the Evolution of Big Data (06 Marks) Explain the characteristics of Big Data, (04 Marks) With a neat block diagram, explain Data Architecture Design (10 Marks) oR Write notes on Analytics Scalability to Big Data and Massive Parallel Processing Platforms. (12 Marks) b. Highlight Big Data Analytics applications with one case study. (08 Marks) Module-2 ‘What are the core components of Hadoop? Explain in brief its each of its components. (10 Marks) Explain Hadoop Distributed File System. (10 Marks) oR Define MapReduce Frame work and its functions. (06 Marks) Write down the steps on the request to MapReduce and the types of process in MapReduce. (10 Marks) Write short notes on Flume Hadoop Tool. (04 Marks) Modu Discuss the characteristics of NoSQL data store along with the features in NoSQL. transactions. (08 Marks) With neat diagrams, explain the following for shared-Nothing Architecture for Big Data Tasks, () Single Server model (ii) Sharding very large databases (iii) Master Slave distribution model (iv) Peer-to-Peer distribution model. (12 Marks) OR Define key-value store with example. What are the advantages of key-value store? (10 Marks) Write down the steps to provide client to read and write values using key-value store. What are the typical uses of key value store? (10 Marks) Module-4 With a neat diagram, explain the process in MapReduce when client submitting a Job, (10 Marks) Explain Hive Integration and work flow steps involved with a diagram, (10 Marks) 1 of210 18CS72, OR Using HiveQL for the following (i) Create a table with partition, ii) Add, rename and drop a partition to a table. (10 Marks) What is PIG in Big Data? Explain the features of PIG. (10 Marks) Module-5 In Machine Learning explain linear and non-linear relationship with essential graphs. (0 Marks) Write the block diagram of text mining process and explain its phases. (10 Marks) OR Define multiple regressions. Write down the examples involved in forecasting and optimization in regression. (10 Marks) Explain the parameters in social graph network topological analysis using centralities and PageRank. (10 Marks) teens 2o0f2Karnetat, Lav society. VishwansHaass bubpanle Pulthde of Technology , Halgyal - $81324 obepactment of Compal Safer + ty sub: Bg Data Analy fie Bubtiode » \6TS42 Ghadp Name 2 Prof. Yasmees: Shatbh Buulioo Papa: + ser BE Exam, Feb (Ass 2022 _Mepy LE = lay| Dagcue Hhe evolubiay o 89 bela * 25|-the atee eo tedimoliqy has led do che prduéleo & shorage of volumt nour anouat of dale - - Carre meg aby ts pee ted bit aswadaye petarylca axe used ‘je proteas ing 7 analy £u 1 dis covering new facts » genuating New browledge . - Conventional Sysichis “for SEK 7 Proterig analjeti pore challenge i Longe Gorin Us elue oy dade, vauely of dat, , Vanlow fous J. fomatr t Fate asing Comple tty j Paoket giacalt, cf-dita og need W quickly prs cerotig, + analy seq re @mpler ¢ Unsruchird PB oda Hy pe ox, relation, lb a 24 “lege, Shuichi Bs Be % ® res E| 68 : i a a Pinipl and Stnechured Auda fypes Sipe ay Complex '1b] Eeplaics He chatacterde tics 3 By bela. | “charackyctetice 5 Bty data cue volume - Beg date coma tum dig ; eAtch aelaltd fesig ot dap 4 Ane the characteishe Sty ejika the anoudt ok quankhy of cate , Which w genentid foro applecatin. ~The Sr3e deleraunts the parcerdin tonstderalions needed jee Aaadleig sthot-dala + ) weloctby ~The tem veloctty refat fo the dpeed genection of data - veleckty tm a Aeadure of Apo Pact she date Generates ¢ protemes. tO Vautety - Br tath — tonprises of a vartety fdata - i -eeata geetoated hom mutt ple So uetes so a tern. — cbc, tonsirts § vortout formu formats « ~atne variety Wo due fs the avafinbtlty ¥ mage umber § hekwogeneous pledfous Yo the im cuuthy. v) venartty — & clo consfdreed ay ay finportanp y chatackedthts 44 ate tote acioudh the qywality of Ante caphired , Iehich Un vaty quctly attecting fo acetate analysty . 1We. > Pandey at neat block albay cary : saplary obaty, AechWechine herrgn . Layers | [ecpors of | [tab weds wag? | |Anvbybe (eat-4 ns : Ae | Ob totlord ie tat ir | [een ea | Bion II [eyeball Layer 4 [pres ebook] Procesting Teil - | ynchronous | Deca 1 Peng Recluse, bre -Sethedated | os | Processing | Hime 4%, part, || lbabehes ox hy | ae | . Jilereay Layers | 3 [donald cattery @ vyte- | [sors ey |Nosar date | fra prnans | Data Renads , Co napreds Shorage ent ies shad heat} v1 | a pee ack cox ore fai oh poe Ara Bs Tahoeatoo | Deda temardic, [Propo coass ‘ngechon of [rai 2a stk ait Exvac| erp area | farce le Uy Ki a bahches Bb aquidog| atk | opr [shone es Pa “Bala Pesca Omar ibe Layers |Soueres jou Pluh feate euapecaaat il gui “| ni ati Rene. | fae evteinal fron the, ifrls, Peboe | eu | leoweres vf date | regen jcete dhe‘ tuber) eat, anclyfhes need -the number eas alepe . fF Bf set, anchStechire dein daoh etuplthér rods aye) the tegical — layear "approach. Prune. above chow s He fogical Lay ev 3 the jrtorm which Ate tonofdurd sry By ala ode elttechure Layer L) Contdire the follosing cupects in a design -fmouct of dati needed at logerhs layer 2 (Le? ~ Push fron Li Se prt by £2 a8 Pet the mecharem foe dhe Mage ~ dourte data - type: obatebaue, yt pod u& fuse ~ Soutte formal te - Cemf-chechired , wrcbuchured 0% Seodhiyed «fly comutdier the Jollors ing aspects: ~Ingtahon 4 ETL processing — eCthee 2p 2eal tru, rehtth means Shre duse Ho Adla cu genta Bled ov to batch , Ls - Courdlees Me 4 ellovw ing oupects: date Srreye type , Joust compromrin , data qnequayy., quetying palteens, g Wrounipler, requivemnt! for L4 ocle, ~ bath oreje waike PPS pc wool dertw res - HBite , CAwandna , Monje DP - 14 conidia the qollowiig aapelts » — ale, provenoing eethoae uch au raapRidste, Hoes PQ, Cpaok 7 GAG poechord 1 pat threanuing — Prive Br weheduled aston o¢ veel ting ov hyditd - ~ prs eenoing Css Syochoonovs @ auyncharnood requiements ab Ls, te=conghduot the tonsunphon of dota Jot te Jolloring ob et integealia, ~ Pata cA weages je An pos hey 4 visualivatis, ~ ainalaru faeal tay neat Real tina , ccheduled both), BPs, Ble ; Rnowkdge dirioneny . —Enpot 9, date. celts jo cloud, web be Or yetern. Caryany | ponte woles 99 Malyttes Ccalabflity 4p ey Hata > prascie pacaltet processing plathrma , 19) pnalyitel gcatabrlity +b Big bata. - Scalabtily enably uesie oc decreate fh the | capacity ot dade, Movage , prs teasing aad analy WFC versrcal scalability — meant scaling up he Qyen syttendi youtees 4 “increasing the aye ters analy , Leporhy vinulisatve 9 capabP lites - Tha & an addtional pay de Colve proHima y qo Conaplbat bits a sealeiy, ap Meant Ausigning vac algeria no actorduy th the archttechire jy Aes OUNCES ay aaa ~ Horfyoatet Sealabrlty Meana faced tee] She number of ayctems norbity in Conhuene 4 sealing ou] the Workload » ~ Procest obéferent dotacey wy a large datacet deploy Aovigortal cealalothty - Wes ~ Sealy out mua wing mote Lmouliey gf as bwliog the pr Leri\y & Chrage rake “Ww parallel - ~The eater tomy fe scale hp 4rale ont exeuileo> of enallt ci U to Peplenenl bigge Machine fry a WIth mote CP foe geet | wlume , volo ctty vadety 4 conplentty rf dicta » &= The — syslem wil pectow bees on a bagev oncchee - However , buys modula + hawlelshe, Jackr 4biyqer meter booed PU be eipensive Conpared ts trtte petfouaance Lachitred by t}ftutar design 4 algoridame « Packs Cpul, Bigger 4 Pass FAs TFoaqscively Parattel protein pleHeemt ~ Scaler tues parateel prceatiha cpicne » ~ Many pre giana ALE late, anol /m coupler yer, eapecalley ty LEP teal compater memory - -Hee Ha required do erhanet (cole) up he Compre Syfiten BL me maroire paraltel provexiy Garp) platform - ~ Pp nnmbrelizatisy pf tack tan be dons oct Mvoral dure 1) Aithtbwlng gepauale sachs oot cepeuls shreacl om te Came CPU it) dtthetbuticy Lopeucc aol tn» copacacte CPI i) bi Same Compicty - W) dishalakiag @epatate taobe ono cepavale Compvilet ~ — naksp le comple ya ources Petvsoing) eybtena « The tompwtcttonal problem tu boeken is duecels piecet 8) Sub-tadl tna tan be pocencd Mauthaneouly Ine eycheny — enecutles tone propia acu a au wel th poraltel cae et anit mona) fal « -ay PBtal -tuc daten nfl be much lets Han bith a Sag Compril, recdutte - i) htepsbrted tomprity model . - A dichtbrlid- computing —mocll, wes Claud, gest cot CluskeU pe ltcs pelien 4 analy x biy 4 lawe Antrutt -6a Archfbutid Comprting — neds oonedd by AS gh- speed nw o- — BPG baty prcercing wilet a pacalta tcolebe a no ~ Ghaag, pogeany anode? , suth as saaphedwe jee conipitaliins oq Ths Hegh Lb hI Big tate Analyds a applecetmn, to PH one Carve shidy, Bly ade anntglicr applecalions cue ) tostetuy 4salu 2» abe leclto) Proud % Bi oh 4 ey Ceecle¢ Rick uRDaG event Alerts Trachog > Healthcare. HY hed Bctin « > Pevactierng > Bigtate Yo aaubibing anal saber -otaka ore Fruperiact jee encue aspect 5 marketing , Sales ¢ advecticvg. aa a analy Hel deplay Lage volune H dela do FdenBhy + Autve Foletttooner Way prdtchie adil about -the ‘tad telunts. 4%y >| bf tile Pralylte in, eledhion cg tanoteliny Prods - Froud Adtercting & Wit to prevent proanctat doses ty sey - — fraud pang lomeone detefving deliberately « ~ BA Late coalyte enadk fraud artectiry - Brjetata Richy ~ Lange Volume and veloctty af 8G data Povielt gerbe metghtr bet alo Auourse alebs uth the deta vied - - ote included may be etaonepoe, tun Aeastate SRL diven § auth a, neo eerees due fy such data. ~ Biy.tatn Cin faim pokatiad Aim fo rditviduall. A company sany cudtee finanetat Losses. 4) B84 acta CueAtt Rich feanagenant —Pinanctal intishdsow , such an banks, extent loan to fadushd ge howehld tecteu Thee tubthbou «4 mrs Counter fee coedt vicky macnly srebe 7) loan ctedaults ti) timely rehan op Kiera g priactpel amount = The Pasig ht tag BG dst, decrees the dtfautl foku us echung § loan , qeeate auecusaly “19 i scitiog ected 4 ~hactke - rdenihrats, § the Rom -payment svftaud ecu 5 the foan reveiving enti tu 4 . 2 TS 284 Seth tAlgonienas Prada ~ Aigmttinne Feadiy fo 0 method of esemtig Vase cele ncihy cubsatid — pr—paguimmedArad’ We bn clidas ACtounting fee Yanably uch | oa bere, petce volume « — Couple Mache water “Con puitAlid mn Tnable algotishnut traded, 4 buimer Ioveadwed — dectcwd fo buy 4celk+ ~The fapnt Anim Ble indghti gatkued hom the ack daoly tes 4 marked dale + ~ Bi dake bryger vlune velocity gvatety w the A redlig provide an edge over othe trading eabtut 6 ) Big Gata FHeatthcare, Bry ata Boaly Kea te Aeattieae we whe follrwsog Aote sonst 7) clinical vetods 1) plurmany re totds WB) elechrna redial retoiy Ww) dragrasty bog # Ark, v) Addthonal datn such aa Arvraltond pom geen baunk aches, medica Leaver are qe?) toufal 8ec DAs + Heahea Grab bor tising Bi date cao fecetetals cutlomu -certue Aeatreaer , pron dtaud , wok, @bue tr healtatur > dusthy 4rvedute health cost, ovens {seen pocktenti t realli ete « 2), Oly Mats to meta Buy Aan — analy Kes dep oy. large volume 9, data, tr identihy § docive de betltgeee wtry predichy modtls about indfvicluals + ~ Big cota dciver cppeacrs Kerp vy teseaeeh in medtcrAe MEL Can Actg pahenta 5 — By cake offers poteatial te han ene medreeds S-heatthcore System> Bfg Rata uA Advabicr — Be kata Hechrolegy 4 analy ko Prowidi rcig Wt Petterns & models 7 Both eelate he media Caqosune rh Atl Contumen ty the put Chase i adsviby of alk Comuners ala. sets ~ nese IM PUL Fiple digital ~ Bu dela hoep h Yoder hy. mangprseal 4 can pode _ ea os ai ened Exec te Late Shady Contdte a drarel agent nebstii ster search meant ee alight behveen ao deafinaley 4 be rehich do nol toanect Une cHly . ‘the seach Shout he geanlte iy oedee a tocreauiry | cay wo thsvugh Stopover at an intermedtcite Aaport © - - The tusheees jane untonPoetabe Loluttens — tre WOME genedled are mechanecal Nitho xT intelligence S$ ophmmigates, - ~The searcdhws shen the che 0 phsons fouct-comeh Ww, Chow wean Such @ Cushorun wed Lead © sthoud. Sleyorte at & ayite & hours - ~The searches -Hherepece neecl ophtmi gators qee parameltes 8) barel cost , amu Feple, Intermed cile Gaporerd & ainleans Yruct WELL provicle martmuony Cuahnue Convernterce as welt Ma LOrt . ~ 8% Ante algovtthons y Advanta Analyser teckriqoy enable pha oplfmryattie, Joc A gquin produ oc Soown & yung ABcrsion «MoDuLe - 2 3a, What are fhe core tomnprnenchi 5 Haclooy 2 Explain brtet- te each 5 ft Conponenbs- A) Pequre below shove Haclooy the (ote Component § The Hakoop tore Components 9 the Franwenort age: y Hatloop tomnrors _ the Commean module tonteery ‘the Uflratrts ¢ uttlflt that ate required by ne OFA modeler 5 +tadooy - fe Hadecp to0mmo provid vastus Components ao inberdatee fot dithtbutid JP Syeteos @ genteat To. Tha facudis Settaltgatidh, Tava RFC FCW bared duty Suche ey Hadoor cofeh butted fie aysten CHPRED, = 8 dara~baoe Ais hibwlid [LU Mytlees Which can shore At Eiki Q data, 0 the lithe at the elusty. By MapReduce vi - sethr ate Progeamacds model ‘in dtedogt Nuina, Mappa fedora. “The Vi fYeltmes toge dataceb op “dala em parnttel fio babchtn: fly YARN sebhoue Manag ing regruriey for compuling The wa applicaks, teaba 0 Mub-taobs aun patalte! "at Ane Hacloor , wrs Sch dulidy bhandle the equals fc he wopuere airhbwted tunring wt the taobs > Mapheduce vi~ ttadoop> YARN based Lgsdemn, fe Poralled procenitg 9 laye datas 4 dtshi bites process a, the appltcaber's tdoba « By Expl Hadoop dtshtbuted Pte Syclear P Hadoep AkchShoted File Syrka CHdeS) ghree the data in a dtch®ented Manner ty order f compute fiust . ~The Ashtbred dota shore 1 HDPE, Lhovr dala to any. format megardert § deme + HDPE profdur Kegh thnugh pah Otter do dala ~cenhoc ay pitedlinrs td wequtre Lage - Cale date Prteraitg Wovbloads - ‘)ipee ata Lovage ~Hecleap datn Store ) Shr data at a ounby of cluske . Each lusty Nas a munvoer oy dety storec, Caled wars + ~Eocth ratk Shores a number 5 date tocle. ~ Loch betantede Mio a laye nuriber 6 clecta blots . ~The xack disttbul acvore aclustee . ~The hedu Rave prolvasing + Srvoge cap alot lites p —The nods fave Yee date io -clate blothe do wey ‘the ap plecatlese, taobi«The data, blechi repleceh by defaubt at lead} on thee Watawodts my came -6t remole nodu, -obata at she Shores enable unnieg she dithfbiles appltcecionm ractuduyy usieg the clusten . B file tondeining tre date divtd indo dada bloc | & data dlock Aefactt! ont @ b4mec fHadeep HDPE Jeahires are A+ Jollons 3) tate , append , dele , Mrame And adtheirt med Recallieg Junctions w) content of todttdual at Caronot oe replated but end > dhe ye Analy Her , data ming, oFP be meditred appe ht neo date at the W) ttt. once bit tend Many Linves during auagel anal proceatity w) Aveage qile Sp Canke mote hyo Some. Pique belo thove He client , master Naneweds Profan Seton danote node end chre Nodes in fine Hasloop grytteal aachGiechire « Clink GQithe wee an the app leealtan nits sac Rep v{ Hadoop ecosytteer prjech > LA singl. Matter node prvide HDF! , Map Reduce , % Hbose ysirg dreads 1 Sma > medium Sige Clune. Wha she Cludle Sit a latge - moullepe Senn aie wed, do balance the load «The Secondary Name wede proviclu Alanunode enunt deoveu 4 eokeeper G Wed by ABase jee metadata shvage ~The anatle Node huodamentally ploy fhe tol of a wordinater - ~ the Mastenode receiver cUert Coane choy , Maintatoe tre duscreptoy 9 tre global jr Cybern anus pate 4 the allocatte § fre blocks Slave Nock ITaakT racker [paiaveds |} |[Detanedte J | | [Paka weet] | [ap Ieeduce | |! ap]] reuse] | precip) [Reauco] The maslee contd’ 9 three Componerly Nanewode , Peondey Nanenede <¢ Job Tracker The = NAMenede Chores att +he ye Sygleor Related ineemahe - erond Alonwnode to atherste 4p. Kanenode « H- fbeepe A copy 4 Mamevod. matadath . Yau Cloed metadata can be vebuett eaurty , b Cove Q Name wecte ote . Whe Yobtnachs Cocdinlt she prentiea proveanty of eee ty: joSefine map Reduee Tramewot6 anol Tu Junctions. P Map edure U A peogiamming enode| fee Ashes biiled Comping 7 Mappee : means Sot hoate foe aovdg Me autianed tosh afte orgdaizuy the data, blocks toporkd wary che beyk- key SpeckRu he Command Lmr of a rep The Command mopl othe bey. fo the deca , hohe an apple cotter, bes. Reducer > meant Lef hoare aid reducing he Mapped dota by uni the agqegaber quay O& tate - spect fied junctios : The seduces provider a Coacive Cohestve wa ponae jee Soe ap plecibeo - | Raqeeqialion uncles, Meany Ye Junctisg Yat groupe the value 44 muthiph moc dgetisr do yuntt a tingle vale 1 mote Segnk he cant neaantag oY meadsuremedl « A fundbto, Luc od Count ,fum, Maalnurs, minimum devtaliog 4 ohh dave dltey « Swenyenay janckion Meant jundlteg shat pre the dutied valve , As prrding foe a but thant of aclacs bho as Shown She bea perfomance to eaaminahdy MopPedre atlone nethrg Application to grocer ably Fhe huge Arne wah oy Ande, 9 paraltel>! jMapledtce unr at pel ausigned Tob by on leuge Clusteu A furt . the clufla fj doex aol Wngl barge ccale dtle analyst ving mudhtpe machina % Yae Clusdet « Write dean the lepc 09 be equat do mapheduee & tre ty pes $] proce) Mapkeduce. Map Reduce pevtdu dwo iro portant junction « [The dichibubis of "job based to cert ap plecdle tak ox weer queey fe valiow nodes Wes a clutte g oc juncios The Strand Junctoo u esq and LAN and reduneny Re waultt freon each node wd a cohesve “wes.pore He tre applecation et answer do the query « J The proceming lasts ate subutited fo the Hadooy: the Hasteop qeamsnooit tn UA Upaager je tab of Tecuiy yey joo enapltish mania) 2 around “tht cluster beroee? dhe dataivedss wih Linc ferp 04 SobT wacker - bdaemon nofus do % gq Aly deds coiled progeam thal background th & spline the whee doe nol contol et partecact beth tals an eamph w Mapkebuce ud Hartoop auytiern « gob yactut , eneculer mun a the ohsch beeps track 6 JAC *jeb swouttted ye noe Tarklracber yee cpacleing dhe tasbes —|Map Kade pegeammshy enables jb souculing tases execution us Jollovse :— A Client psc butt & Mqulad of ap appear do he JobI racke » ~ A opTracker be a ftecloop Aaemon« The follow ade the Sheps 09 the requut do Map Reduce E71 elf mat. toc need 03 nogoure es foe prteing Ahad Tequet TH) analyye tne stale oy the Slave nodes BD place the mapping tabi & queve. NV) monitor the prrgure 64 faat § on the jat lure, matact Yhe task oo se 5 Ine arvatleble - The fob executsy 8 conhothed by Avo fy pes 9, comes C0 Redince ! i, Ramm +4 map -fasba 09 the wlods ~ Map arbre ago to there nodes White the data dpe the applécalion b Shored - The Reducer ovct pt droanidett to the cleinrb note Ole the Anta cerfaligeciern htinq Aver, at. The Hastoop yltum og end the panp + Pe elute “Jobe do the appre pefate dant tathe clus ~The Hakoop jMawvoe £0 hea maaang ee the fouk -of Vesuing yee job comp Lelia $ wopyay dati armunrd ~e cluta “behdeer the glare nodes. ~ Fray 7 the Clute Collects 4 recluces _ The dake +o obs ie veontt 4 wend Pr back the Fakbop Seaver after tonaplilten dy The vey fasts + teAG The Web axcuultey 4 forolleA by doo ru 4 poteme ah ranp Redute . : ~A Single inate prota cabled JasTracher U one. ~The proce Coordinaly alt Job! aunnig dhe Clute + aaign wap 4 redive tasbs do 10 On he Touk Tyarkeas. ~The Letwnd be a mundoer @, subordinate proceed Called Tam Trackeg . — Thee proceais Fun aacigned tabs & pete decally Trepp he prsgres dy He FobTracker. poh Sheet not oy Plume +lacloop ‘ool. > | Apache flume provides OO AlelsPouted , weleeue 4+ avallabe feovice , Flume ey Eck erctly colvects > aqqegata rand transfers A lage amoud 8 cheamig cata inbo HUPS . L Plume enables upload sy lage frlu int Hadoop clustes- Feoliier 4 flume fndude robudoes HY autt pletonee. Plume provider data tranche whech & Leldobe + provide Tecorery ww eae 6) Jarlure, t Flume é& weJnl ~Jor Franateucg a latge ammount 3 dan & ap ple cattone pelatea b loge nedw orb drafty, Senso dala , geo -Woeatis, alate. , enafl COreML Wed — MeCaged+ pApathe flume das jour Components T) Bounce ~ pontels actept doda ism Cetra 9 9 epplicahosWSfak- heh rovewe date fibre ad io HDR rept lou) e Joanmtt the aati do anther Lourle . Wi) Channels - coonect beboeer (oute 4 eink by queving erent late tor -pansactions. N) Agent + tun tthe Cintt 4 Courter 1h Flume, _PAoDvLE- 3 Sa. |DPccum the charactertstea lon wh the feahtrer in ocak drantactions. > | chatacterfcfict 4 Necot date chore DY Migh 4 eary s colabrlity LNecar daly Gor are designed oo ep and Aort yontally- Py Sup post doreplecabion pantheple copies 6 doco. Chre AC WCC TOL ple nro leg 44 Clusters TA try uve hegh avatlabelity , pactrhra , seltabrly + faurt frlerance. » attehetbuteble; gig-stat, Souibére permet chardig atelntbubion a shad on muulhiple elureas woah enhancer perfomance ~& Hovsugh pil i Dd wage a Nodal ceevtw tshtch ate lex capennie tj) Nocal dal, Stove _ we car dat Chu neque here manage mont eps « Tt supyrd feahuer ltbe repari , ease cata dbalton 4 ple data wodels rar mates Aotabase adminis rotor & tinny requivenert Lens hafngedt - F> lager 4 ope -coutce hol? Neel datk Lhe ae cheap J dyer courte - » entaboue ~iruplementlcin & Wy + hyppeally wey cheap ceavert 4 manage the erpleding eta g 1aTranathions hil RDGML databases ar Crpentwe 4p use by Can 4 Shyage oyidemt > Suppoot ts Sdumalex aadamedul: - data con be jeseted 65) mogar date Sow jofthoul aay predjer Schoma * F> Swpport te “nkgpated caching # Thu thucates Cyflem pujemance. car databate nee separate Tofrathuchire fe that No fojearbiity ualtbe_ she far [Dem 2 wecal re ate Rar ome oe more 6} He ACID prepeshts ®) Chacala ge by 2 entof? propertion 9 Cap Mheoeem WO Can be Chanacksi zed by Bare propestier - “|htt nem Aragcan , uplain she potlowiag fev Phused~ Hhotheag echskechure fpr BA Dela Jauds » 1) Seg ~cenm mode 0) Sheedy yorylauye dectabetes i) Mastic clave modeo Ww) Peeo do pete modot rr) Stage Leerec model Applechin veqdired Aatoseti cléetrtbild Per instance Single — Ger vee Dunning the apple cation- Fu the civ plat Alechtubil, ophers joc Nocoe dat Yor . ~ A qoph dectabare processes the relationt hep! bebveer neces ect 1 Cetver. - The SSD modt Get jolts fe gaph PBe. ~ Aqgre gate Fi dotacets be key valve ,' column ~ Pautly ot Big Table dot, Sos Which require Ceqvertial procuing - - Thue clata Shover ako we “the wep model, — tp applecelion excuiler the data eequondtally 09 w& Singer res | pe ven losge soatacets Sexvet Bk th a clurter Fegure above ehorvr Chardeng ot very. Luge datasets Yoto Jour AbsctOns, cach NenNUy the applecdiies aw fou dbheent sures f, 5,4 60 at the cluster b Der, pe; , PBR «dB, me pou Chard . Exetute Overy + tte Poder face ( clr v7 Web ioteface) Senda a query fo @atabare datver to exenctt the avery - L- Get Plan * defve derda the query to query couptler that pau the quesy de Check the Lyntae + awe plan oe the wequicemant % The query. P Get tactadata: compte Sends metadata request fo meta chore ~kend metadata: taeladore ctencts metadata au A wapore fo Comptley « Sond Plan Comptes checks Ihe gequtromedt andl veaendu sine plan 4a Aime. The parsing 4coupiling Sh qwey templet at the law. 2g6. Exewte Plan! obetrn tends the vexeuile plan be Meude y ngne: 4 Evewds Job? arlenatly, the prces 9 exetubion Jeu a nap Redirce, "yoh fends the ob te Fob Tracte . whrch U in Namenode Gi Ad Gn thin yob to Tak tracky Ahh ty 1 Bate Nede, “Ther Yhe qvey exeusb the pb B. factaddla operation : tenn lode vhe exoctier’, enyre Can exotile the metadata operations rh metas lore - 4. fetch Ret Cxeuilron engin Deceiver “the wears yoo Wao Neder 10 Lend Reanlte: Exeutlion enaini —genols the seoutt - 49 Dever He Cond Reals : Debrer ends dye vepnlk fo ts tobetater . vege the Yolloror wi) hao , rename +d Oop O postthon to a table. 7 |F) Geate a tabhe VFB partihen, qotlowing Command YU wed bh creat a tabi Wirt partshon CReAéte Lewtetvacd table Kcolumn Mame £7 C lonamenT App Cap now Exzeig] PARTITION —partihra cree Ctocataon ‘location ] Pasthmn —cpee Elotarron 'locahen2') . = Cn: AtTER TABLE fey tbl App FALTIT ION C caktgony =" Toy Aree lane’) letahon "7 Toy - Ars plane /poot Moplane! ) ®) rename apatithon 19 the emishng table tuting dhe qollowsing cornmand ALTER Taos < tablenames PARTIT TON Pas Wh 00 -cpec fenamie TP FAtTIT LON Pasithio - bRop CIF Ew rex] PARTITION farWhoo-cpec , PARTETIO~ parhhonges ent ALTER TeBlLe toy Hol DROP Cap ewzere] PARTITION Ceategory at Toy. — Atop lane 7 8B) iohat & P24 io Big daly } typlai the dela Q P2G. 7 |Bpache ereloped PEG , whtch: _ Fe an obshactiay over raapheduce - &o an exetuhty Pecrcewort oe parallel processiq ~ reduces the conpleatleey 5 tothag a facipRedlurce Progylans - Gia Aagh-leret data tor oy ucage. - ~u mocHy wed ts Hoes envirdomerd » ~ ea jeues ata, Trantp wlalio', Operettions at {rls ot olata nedu th Hadpop. | odes ms a ? Wy bpache PLE Aedpe Programme torite comp lea dota tran foemation ang corp pig Leder longuage a ver Stutla 4o sor ¢ Posse’ QQ etch ee} brfU mtn operaley , Such a Georg, Join, fiber , Dene , otder by | Patalted , cod & spltt ~ Progtammens — wetty Scr pds wring Pig Late, fo analyze cata . 3}“The peakpty oe Irvdeenally converted fe Map + Ledue sabe wPth the hetp ofthe Component Known ar txennliia Engire , what aceeplt the Pra. Lab POtpa wy fopat ¢ wnveth Awe covey job Map Reduse “jobe « i) Creates = ea ee Ae} thed junckons Cuppe) to wate culdomee funchtdne tshFch ate not avallabk iy Pig . ~A& Ye can be eH other progpamnti Languages, Buh a Jara, Python, Kuby , Jy Mon , Tktby » They eautly embed font Pg scat ndtien ‘9 Peg lates - Tt) Peocers any kind 4 data, Swrechued , Sent - Stouchued a urthnechued date , Coniing from Voriour Sources + I) Redtetes the lng of codes tuuing maf - query ” aaa Pty Coole gto Liner ia equal +> map Reoleece Code 4 00 deiy , Thus "Processing fe veey fas) V) Handles ipeonsittat schema v5 Case 9 une hrechured data ae wells ) Exact the dette, pero opewations om trait dock & dunpe the dada th) the naquived joemat in HDPE. The operation te caved ETL Uptract Trontdoune lead ) 82jv) Pea forma aubonutec o pHonization y taobe Afpore ewccultoa - iit) Peogeammets ¢ develope can contentrate 09 the rohole. opetales., niftert a recd to urea mappes 4 reducer ~faobr Separately. ww) Reade he Yap data {ilu from Hbps or Fe clita Giles JOO Otee Souter Luh lo cat Yl Salem, Shree: te fotenedéeate data 4 wwhy bade Be owtpri 1A HDFE. *) Pq chauaclerisher ate rate e0eaduy , proveasng , Pregeaninuing We vats in muttspe Languages & Progamming murtep queriet by femer coder y | This caves pani Proteaacag. : xi) Pq desire Guidance fron jowe phelovophi'es, Lie anyiohure , take any teeg , domestic ¢ nin as ae This Gubbis the nae piq, as “he antaal 1°94 alto Aas Saree charachecche - Mopvis -© 2m taachine Learning crplain Qinear ¢ non-Lerweae Relachinsht p hhh ekKenteal aeph Reon selotsonch b |- 4 Linear adostironshep exist bebveen 90 vartable, say % dy, RED A S hrakght Unie CY? 4 #41, 2) can Je oj a geeph, PF atleast some veauonabe | 32dentec of acturaty, Th a, @ He Uneartty Coel{t cient. ~ Examph: a ccatter Chath Can suggest a bingas Relactia ns hit , wAtch means a Chrakgnt Lure . Ls Pegure below Shows a ccottee plat, lhech ie a Lins — telabonchip bebveess re number vf Shudente optuny for compute Courter oh year bebween 2000 € tolt . 2000 2005 2olo ole 2020 eax —A Lorne Aoladioo skis can be negahre oS fost hve « ~ A posthve vetoborship implies 7 Oo vartable Taueater fy vale dpe t wolue, other alio increques WO -4 neqatwe mlabbochip p Puplee , whin one vector Pnuearer —') value, the Oer decreaies ih value. ~ Perfect, ehong . oc weak ineat crip categuate chepernk upon she bonded bebveen the vo varfablu . 34| Alon. Uoean te-labioo shy ee eer sLetocion chip te Lakh do exist behween uo gpuank-fochv varsabke olin a Cuove Y= aot Air taare...y can be uted to 44 the data poinds . | AO Apprcahty [Phates + Tex} Pre Pool etirg, By enables gyntahe | cemanter deel = analyec ¢ docs the {otloroing « ‘ed eo} Cleanup — pro ter 5 removing wunne ced aly BE wn wanted th Sy 4 Tobenigalien — procem 9 fplhy he cleanup kewl ado 4obent- va ira) vehffe pater punthiahon maobs >> Pout 0 cyeech Tagguig - Por tagguig t a metrod thot cHenpts label sy each faken sith a Ay proy-ttacte A fa conse Atsambt guahon ~ methed. Whech FdeolSher w- forte § weed raed eb g sentence, that ‘gig ” blaring case Ye wood cher onnttyple meantigt . a> Posing ba mutthod eich genteater a paye bree dor each fendener, Parsing aHemptt finger the preaze qeammadscal oloKonthyp beeen Affewnd woods lo a Cenlence » 36| Phase a: Reabure_Gunsshi ty a proves Whgeh yd doy nee Feahues , geome ty the tag john genradion awe Bag oy words - Text Ascumed ey vvog reacted. by word Me containg. ofacunend Clanifcahdbs metrody - eoumonly we bag~el-rspacls model» a Stemnuing = thant weed by Pu toot » Ty rediuccet svord de tr med baste chmect. en senpuriBeahio = pure Rancovidg bop words frond te feohure Spates - Hoy Aree COmnuon woods, Ualfkely fo hetp terol miaity - Wy~ vecte Lpac moder Cyn) — cn aby ebraic model ype repweoenting Hb docuncent as verter 4 \derbheer, toowd Jrequenciée tC Hema i the document fndia « Phase 2: Feahuee Selechin : - & the procen thal delects a Subret 9 fevbue by rejeclirg juelevant and /oe sedundand peahve: ~ TA pesjouus Aimenci oaatity teducktoh , w-oeane evaluatiey wode —— aan .. - Feahue (elecltd, AyrAihny -reduter —cbnaensrOnalet thot aot Only Inprre Ac pepomeanes q basnteg I wlanrdirn bat Alin radutes dhe Shorage reg Ure ning fot a clatacet » ~The protien erhances decte, unduttanding § visuatigahsg hase babe Auuning deehniques F -TAL enable Tautght abou he Shuchued database Saat reanbted ftom he previout Phases «| -2t 49 wolws WrLuperw ced barring Cex clusteteng Uy supeviced: teapning Cex clanisrcabioy 9., Bde ‘ evo lurhionaty pacttetns In Femporal eet Groans hase analysing vest li 7) evaluate dre ontime a the compile Pro ens 1) nteprefalic, 9, teaubt —dt a Cepia Ren meant oblduud car be ded aa Qn Saprd po ment cet Sequencel Ele , the wewtt-can be discarded + by b Undustand what 4 tohy the potun 4okhed « Ww) Viswaligabids ~ prepare vitual pom dala | butldA a prlohy pe W) ate The rasults Joe furtice siprorement adler a4 the exlupate Yodushy. ot inchthution « toa) poepre mlhip le reqreciont nette down the spl involved jeeccaating anc ophimlg choo ee vagresTOn « Murlpl, vegretfors are wed then Tode pendent joctou ate tayolved. | these wragrenions are also widely tued do mabe Shoe- 4p sd ofeear predictions do Aen Whech jacoy dotodude + Which do eacluch . f Mutheple sogcerstons can be wed fo Aurela denn le mode with ditfecent Jachu Ayo et exyere P More than one vattabk canbe wed ay prodicéloe WA enuttsple TOgreAstone .The: prealfthin tabu the yom Y= A+C A+ Cran Fe... la Any where A L the jpkecep) -G the luxe 00 the y ake C102, Cn We Coe}) Rélen , reprecentiog The Comtotbubiay 9 the fodependert varbables DV, %2,---20A ® the caleutatrs, 1 Sho The cramps Sovalved 1 forecathyy and Sp Hfonigchoy bo TegAEAL Cory are 2 Using kent analyst oo fale olotn EH monthly Sales a company. Contd Porecaut 8 ala je jodve mooshs - W) fot tae quads Fru A Company Aas jonveted ©] tmaakelidg 0 parlteular brand, ap analyttr 4 Wheater the daveduend Aaa gf Subttandsal whune or ro} Cen be made, Rt) Suppose wo pronobion, ~Coupasgry Ate punning © Ty , Roots vw parallel. 4 hitnr geqiertion Can Con}-At the Individual as welbacthe Combrned Tega q sanmuag ~frese gavordscenrents dog char » ) dp waveance — Compal erploty a LA eat Legreadtin model to obi a centabey Plerclun fable dling predécted claims fo Tnemted steclared valve talfo, ) A ffrancat Compa, mary be *-Fereated bo mui entgtng FH pick poolfolio & Rance tant fo andurtdand “the top Jive Jorltu gt tearons dor default by a Cor hous. 34Vi) TM predict the Chuacteehy a chitd baced co he character ou 9, Huth parents - Mii) A Companys pac an employ nemt dloictt niin adyen matte ad Which @& clatmy that Nonun ar betng Abscl wtoabed agatnet fo Jeu § calouy uw Aaiced . i) Predtcling dre pret mq howe , constdertig the locality g build charadsisha wm alemliby 9 + echntg ues i x) Chndsng Lotationchtps bedveen the trichue § the Bologreat achryiby 6] Compounds through Hai phyfreal , eet 4 phyechinancad batty mor} Commonly pee WP Mogerawy fethairqve |x) Yo predlect Compound 1H Nihon bre aetevity bits grOupl « tedp) Snpluin pagamertees In Socal glaph néhwor k fopologrcal onalycs tay Cental tlece PageRank, >| Paameler to coctal Faph neksosk fopoleg Fal analyse taieg Centralttier 4 Poge Pork ane 7 obegroe ~beqiee 0 A gaph verter Means he edges linbed te trod -In-d ameter rand the number 1 id -cdges bron the GHhan reedites ~ Oud dogree IMeank the nunaber 4 on eg es qrom tad euler fotal nbeubee|2) cloceness Graph rede Closeness ¢(v) a way che fg he tireably Hoa weltr & whence to her verts ty + ~The centrality ju Punchin 1) aivtaney of Ver pias +t ow) =[2 Aus} uév ney Whe Alu) te dictane bebreen Ut V jew foe pact ara at. Poe) ex tne acee eyjechve clorenes Cgey) Ean abo be analysed, we appedimade avecoge Aftlance Jr Vv fo alt ome veriteu to plate wf the shostery paky . Cec Mantes rn Lene pr carer with alae Number oh edges + nage Lata Scalabtiédy — ty Compualons 4) Bobseennest Graph wrbces befwennen Means The Dumbet 4 Iau vouka tated y bobwern pe Shovtat po y the valent do which 4 vedlen & tocalid Ledween other pair vertf ce . ae kak Page Ranh 1 Anutne por the oolance of earh vwelr an aA gage Orsuming an edge pom Vi devin represents tndouement 9 Importance Q V2 byw, by connetling jpoton’g , iottracleng 7 ophag fpr culatroniny , Sharing belied o. Some dpe mutant » 4}op ntacts Sie Contacts Sy meta Avetter loonection dy many ver}f Us « HS tndizeck Comactt Ondvvect Contac mehse prinans betweenness , which de the sane § the Shotkeatt pals mitted geodearc Aistances fom alt othe, faintng werhtaas . ~ Thee ~Step contact = Mifta Meant gq number § edajes fo other neckteu peu dhe number 4 edgen from Rue roche pha geodesic Aish g=3. by Struchto obiverst ty Shruchure. Abvenrly Roan: oe artes ty Aereete Sie —geaphe. Day Facathy flame 4 Signahure Gx Dean Beatend cs Prof. Yasucen. Shatbs chee Co wkad guaph hho A