[x Stasstical Measures applicable t- Eno-Ter EXAMEN ATION _ c at CJ 4 Scatssg Data. Statistical Measures Api cable on them. [a Major Scades, 4. Dato. eb Nominal Data Reprise cab Staristi cal Measures styucen Re Je Mode — Atghest 2+ Prequeouy a distsibuten— count of cosewarion wiyein each 7 dD Ordinal Pat ane craved ina meaninp hut sequence gt O Yank. Repsesents careforical dara where values [1s Median - Middte most value. Divides datainte # ial bons, 2. Quartile ~ Divides data into 4 posts: Cequald 3. ferceatile - Divides data into woo equal parts. "|Utt®) aatervat Scake = Represent data where the Aijercace perwern two values #3 meaning | + These is no true tero fofat tn thfs scale. Cate) + Tadudes both ponttive and nefatve Valies # Statstical Measures applicablet- | \+ Mean - Average 4 Valtses : 2. Standard Deviation - arnount: of Navi about mean - 2. Range ¢ dijyrance between Nighest and lowest means: 4. Variance - spread belrotem valird S- Median - middle most values. 6. Mode = Hrghate yrequeatiam - ‘ p wv) Ratio & 1 Repsaents dota wherethe watto beween dota a fotnts |valuss fe TORRE I. Seat) pate hos o vue zero polnb Scanned with CamScannerSS (ee a eee Lud dddé + Fncludes only postttve- x Staristical Measures abplicaile 1. Standavd Deviation - amount | yamiahion about mean fg. Mean Avnvage e atl values 3. Coufraent 4 aviation = velative variability qf data. axound mean @& Explato the preproceming: tectniquis 4 Standardisatfon and Novmatisation | co) Outiiers Identification and ways to deal with outliers. Alcoa and Normatisaton ave both techniques Caed fn data pre-proceratn& to scale numerical jeatures or Navfables to asimilay range. # Beth axe sfmilay fn the foal i scalin§ Aaka ALljyen tn approach, ate Apa disetibutten G daka. >| Rescatas the data 40 that t > scating- (a) 2 ft has a mean 4 '0' and }« standard dewiation 9 ‘1’, Sener + Centers the data around mean and scales tt by standard dowtatton ib Scanned with CamScanner- [Rasnetfeatien meee Value me (03/84 Wednesda ~ v = ming =r ty era ay Caries data sothabs E) Novmakicatt Rescates th, Data falls into a vanfe between “ot and “4" ~Hecarty scales the data based on manddmum and mtatmuen values” aatatrum, sNormauisatten ean be AUT Eee cree and martinum ave abpecteat by thorn era doa oo ee Tdentityiah anc Gating missing valus fs cxuctal to data 12 eneure aectvake and vellable vesuits- : O Visual zospeckton - Niewadisin§ the data using chats or “ sed missing values, Gaps tn plot or charts Indicate missing data botnts (2 Summary Statistics = Coteudating summary stab'chtcs Such as Mean, median or count can eveab tatsstnge Vatues ba Unusual vales Such ob extreme outliers 2 unexpected | freqtenctes , might fodteate ma tsstng data- 4 _b flegraromtng methods > Usiliain prebrommte ta fuafes fuchas Fyinon oF Rand using thety Speethic piactions to Askeot mfesind aliens, : if B)dernato tnowleser UndsrstandtoG We coarouk and tts expitted Values can hel Beaty niesin, _ doko. Condtttonal formating “9 ME Excel Hkertag Count Blank function Scanned with CamScanner3/22/24 , Wednes day o i Ferment g Mesto Values t= ~_@ deeeetens Delete vartabies| records welth MteshnR Malis. Tis method is Shralghtieroard but may Lad to tose 4 valuable ¢ t _Inormatten, especially tt nfesing values ase not ander. “T's may afyeut tne Exealofity and auuthentidty of the data post analysis. " lacement + Replace méssinf values cvith estmated or coludabed Values Mean zi no outlters wctet. Medan — ft] curlers eetst. | @ Using mere advanced techniques Dive Regrensten Fmputation E-neavest nelghboure Imputation Ck -nn} @) Missing values can also be replaced | fe using the iW wp down obticn zu —— vi S Gi) Frediictten Models’: Prediction models can be wisedl to fill missing values provided the nofssing values are In tanget Nartable. Predichien {5 bared en alternate voviables. __ ©) Advanced Techniquia Mutttble fmputatton techniquus that wikilise algorttnins to fenewake data fer mfasing | blank Spaces such arty @ Dectston Tee @ fandem forest Scanned with CamScanner| _ "stu _ ag _|__ performance and usabs fa tes pred afgroach eeeesiatinne eb fcor (a) Bxplatn Stay hema to Power BE with an example. (0) Fuplatn Fact and Simensien Tables 2 60 Explaia the Ailjrence between Normalised and - Denormaltted data Oe tay ie 6 Explain the tured case 4 Stax Schema y 5 Sovofiake schema 600 Stoy schema {5 avemmen dater AHodelling technique used — fer erganisin§ data into a c4nbrol ject table Ssuxroundecl YY Uiroonsion tables — Ceales oy rover) + The eee table cor tans numerical Mesures, While dimension tables Provide centent for those | ike preduct, autemer, or time | __+ This dustin Simplbjes: quenin§ and an __Nfsualteattons. cacti * 4 to Yodarstand and comprehend 2 Met OF Cousins tarnifence’) teds opFintse. use tprth’s Schema ee PS it _ Stax Schema ———. “Dake =D a Date Ohvensten sme, tke ———— a srr Product pimension ¥ Proctace r)-——— aa ea — tabeGor i shee > Location, “asd —_______* fete ee + Supplier * tustorec®), oe +Totalsats + Qty sad Scanned with CamScannersuave © bsbad 2a, See . |e) Fact Tables i= 1A fect table eenesas a ceotral table in data medietting techniques suchas Stor Schema and Snexa| schema. -A face table dypicathy Contains quanti tattve data, alee tf Pnown as measures OF metres that yebsesent dui nuns transactions oF events » Cadlis Seuenu, Prefits, quantity sold). + Graqulaityr aut tables ave usually ok a transachion oh bueel 4 franulart 2 Capturing individuial events g transachons : eI Powe BIt Crucial tor creating dnsi reports ae: in Fousew = stati Nurmevic data tan be On a fared on Vomious dimensions Orie ee Steed in Daj pal dirmensven tables | | Disnenstem Tables + A Dimension Table Plays a erurtal rete io oxgani sing [reviding emtext to te data stexed tn feet table. +A Dimension table contains qyalttatine [dsrent tive. attributes that prowide context anol additional data in the {alt takte- Customer 0, automey tocitien, cademer < encle + Hetp fn Cateforvaing analys ais and fpitestog data ao “vanfous pespecttues. bebe Py UT nr Scanned with CamScanner~ Structuye nae Yelated. tat —Mivolves fever tables Sombe a ‘ a ao Bora Mock pcatibg, sila Rexel to pe ré 1 pad —t | I ing desta integrity. | Atembromvses! , data Inte, —— ams eas as og _[t Mest Br tocls eptionise use Icey Nemalised Dako ____—DeNesmnaticed Data _ Procese 4 exganising decba to Process Lamers fables enininnse vudiiedany ard to reducce tne Se Sl Sang by ANiding large eins Nscded for queries, often 4 | fables into Smatler ense and __Tautking tn duplicated data, fedefining owlationships betwee but faster tuemy besformanee. fin Y fpr query be Rormatised dato te stosed in Penevmalised data usual __Peqpives move storage space ____ dhe to Aupicared a dato across A Sei ss tables Cauplicated dasel) =a enfin need te be. ry hy alo nea od matitbe Places Com lene oa «pins Foster g : Clovsan ee with a oe " large datas) a. lerve © Opdates |roed it Feahiens | (A) Star-Schema #Propwredabpreach = Tae MEAP ASAE Corn _ft Fasy to understand —spreeath ders used » Jd If seme data woanehs fpr this eche Oakes WIP many -te-ma > Fowes & Br dvab of-and ~avop a S hetroten dimensim joo pole a pp Sed choice fer smaltecdate pe Supper jira ‘sues, fe Comper quesrits Find multiple dimension tables Scanned with CamScanner