《米其林的数据民主化.pdf》由会员分享，可在线阅读，更多相关《米其林的数据民主化.pdf（40页珍藏版）》请在三个皮匠报告上搜索。

1、1DATA+AI SUMMIT 2023 San Francisco06/27/2023234Who are we?SmartAutomation&DigitalTransformerIT Expert on Data&Analytics567891011HOW DO WE BECOME DATA DRIVEN?12Data Platforms13The Corporate DatalakeData visualizationProcess MiningFT AAAFT BBBSelf ServiceDATA GOVERNANCE/DATA QUALITY14NICE TO MEET YOU!

2、Support functions63 400FTEsDOCBSa worldwide community of servicesIFINANCEO2C&LOGPERSONNELPURCHASINGBusinessProcessManagementData office15Customer Contact CenterInvoice to CashCentre dExcellenceACTIVITSOriginal Equipment(OE)Replacement(RT)Business support Services&SolutionsVentes diversMARCHSLOCALISA

3、TIONS9 Customer Contact Centers(7 RT&2 OE),1 Consumer Care,1 Shared Service Center&1 Center dExcellence.820 EmploysPROCESSUS O2CCHIFFRES CLSORDER TO CASH:at a glance!2,500,000 Lignes de commandes700,000 Appels traits50,000 Rclamations2,500,000 Factures Value of Data for infrastream usageNeed for dat

4、a and automated reporting pulled by lean managementUsage of Data is very manual with excel as primary toolSharing of Data/Usage of shared Data:Mix of data store and extraction of data in live application.Too few Data in CDL¬ used due to bad experiences.Value of shared Data across silo is understo

5、odNeed for data and automated reporting to enhance customer service Usage of Data is enhanced through availability in lake,state of the art preparation tool and data visualizationRecurrent reports are refreshed automaticallyData of the stream is shared in CDL and combined to data from otherstreams f

6、or more insights.Allow first experimentation with data science,AIO2C Data drivenAS ISHistoric industrialCompanyTO BEData as an assetNext StepData driven Leverage machine learning to support complex decision Empowered,augmented human beingSimplifier/rduire la quantit de textePropose best option in ca

7、se of unavailabilityDetect and solve issue order executionflowAnticipate cancelationData for customer service improvementOrderCapturePortfolio improvementImprove councelling during order captureMichelin price vs competition priceTarget vs.Achievement for given offerLack of availability of competitio

8、nAvailable alternative for non available tiresOrder portfolio historyEvolution of availability and priceson the marketPropose targeted simplification Cost to serveCost to serve for practices vs customer vs turnover generatedTargeted improvementData requiredKey examples garder?PerimeterO2C&LOGEuropeR

9、emplacement(RT)(RE)Understand core DataEngineer rich dataset Automatize the reportingVersatility proof LOGISTIQUE Europe+IntercontinentalAutomatize reportingFacilitate diagnosis2 million order lines 30%of group turnover300M/yearKeep factories runningManage crisis90%10%5%95%TodayTomorrowAutomatize re

10、portingDeliver added value with insights provided the efficient wayAnalyse,share and decide Manual and repetitive tasks to reproduce recurrent reportingExpectationsBe GREAT at BASICSPrescriptive AnalysisPredictive AnalysisDescriptive AnalysisCurated DataExposed DataSupport steering with automaticall

11、y updated data Deliver insights for strategic decisions,validate projectsHow to?Automate recurrentChartsFacilitateAdhoc anal yticsSupport experimentationbased on DataData available in the cloud,ready to be usedMain componentsCentralizedstorage of all usefull dataOne reference data sourceData as“raw”

12、as possibleIdeally keep Data scheme:Doc of SI=Doc of stored DataAutomatic updAteof reCurrentreportAutomated refresh of recurrent Dashboard/Charts/Indicators OrganizationIdentifyorganization to deliver insightsIn traditional tools,data visualization build by IS/ITExperimentations show opportunity to

13、leverage business experts trained on data visualization CorporateData lakeMakedata availableISBuild data visualizationQuery and structure dataBZ+ISBZ+ISDelivrablesof the project teamBuild:process,actors,toolsExtraction of DataPreparationof DatasetPreparation of data visualizationand AnalyticsCreatio

14、n/Update of records provided by business processInsight,Decisions,ImprovementsImprovedsituation100%IS/IT0%BZ50%IS/IT50%BZ10%IS/IT90%BZExtract/Publish:Attunity/KafkaStorage Raw Data:Azure CorporateData LakeData Preparation:DatabricksPrepared Data Storage:Azure Corporate Data lakeFinetune Data Prepara

15、tion:DatabricksData Vizualisation:Power BISource(Bronze)Agregated(Silver)Fit for Purpose(Gold)Key learningsFear of the prompt can be beatenBusiness can create scripts(fast)Business need IT to make scripts betterCollaboration is key for sustainabilityHOW TO SUPPORT DATA CITIZEN JOURNEY?CitizenWhat is

16、collibra?Data catalog?E-learning CollibraIdentify the good datasetVu que cest D2,demander AccsContacter le Data OwnerRequestAccess in Portail User autonomySharepointCDLDatabricksSelf ServiceElearningTraining PowerBIRequest Pro Licence Create yourworkspaceFaire Passer son workspacepremiumPublishIt is

17、 not up to date!?RefreshautomaticallyA not so easy PATH.Success Storybased on real eventsBack in 2020 La captation dune partie de la data se fait en automatique avec API ou autre.Rapports PBI automatiss pour reprsenter des volutions de prix des matires premires et prix marche de commodits,volution d

18、e capacits de production,analyse de comptitivit prix dachat vs prix march etcAutonomie dutilisation CDL,utilisation des rapports PBI assez lev.Le besoin volue dans le sens dintgration de plus de data sur des domaines diffrentes nergie,logistique par zone analyse dvolution+corrlation par zone.Probabi

19、lit de revenir 100%?Support,MentoringGet some Fresh air,Out of the BOX thinkingBenefits for both sidesGrow Up on DATA TechsBetter Understand IT World and ConstraintsCan Reshare her/his knowledgeWITH COLLEAGUESBetter Understand BusinessAnd why we do what we doWIN-WIN RELATIONSHIP32Auteur/Dp:Date de c

20、ration:Classification:D3Conservation:WARf./Document:THANK YOU!We are part of Data governance processExposeWhatDefinePublishDocumentGive accessManage quaityCretate diret flows to CDL StoreWhat:Provide relevant history Data catalogue Relevant datamart to gain access,value,performance Extract ValueWhat

21、:Build Data applications to help SC performance,efficiency,sustainability As contributors to projectsAs leader of data projectsHelp others to extract valueWhat:Spread data usage across the organisation and position Ease access to specific data module(data mart or data block)Facilitate WW BI/data net

22、work COE Data and Analytics is here BIAIA3To put in place the foundation of further analysis and show clearly what is the past and current performance.CTo understand performance with a holisticview,todetectactionlevers,toactautomatically locally and globally.4 activities(for the moment)To help many

23、users to useand get value for our data.To help to predict what willhappen and what to do.Plan pour aligner lorganisation Check pour vrifier si le plan se ralise Act pour adapter le plan la ralit Plan Check ActDecisionPlan Check ActPlan Check ActPlan Check ActPlan Check ActPlan Check ActPlan Check Ac

24、tLA SC cest Lart de faire chaque semaine un plan faux sur des donnees fausses en toute bonne foisEt de recommencerPlan pour aligner lorganisation Controle pour vrifier si le plan se ralise Action pour adapter le plan la ralit Plan .COE DNA RANGE OF ACTIVITIES FROM BI TO AI THRU ACTIONABLE ADVANCE AN

25、ALYTICS AAA Ai projectsBI projects Advanced Actionable analytics KPI&ReportsDashboard BIAICAD 123AAA:automatic root cause analysis AAA projetHuman Local actionAutomatic Local actionProbalistic predictionRecommandation of action Automatic action Data miningUnderstandUnderstandFaster Lucidity Act glob

26、allyAct locally with human when possibleLucidityUnderstand globallyAct locally Act locally massively UnderstandDeeperEfficienceSC Performance Decision plus rapide et plus AutonomeProcess Mining Ai projectsBI projects Advanced Actionable analytics KPI&ReportsDashboard BIAICAD 123AAA:automatic root ca

27、use analysis AAA projetHuman Local actionAutomatic Local actionProbalistic predictionRecommandation of action Automatic action Data miningUnderstandUnderstandFaster LucidityLucidityAct locally and globally Act locally massively UnderstandDeeperEfficienceSC Performance Decision plus rapide et plus Au

28、tonomePlant Michelin DC CustomerForecast Plant Stock Order not fullfilledIs the orderadnormal?Is the Safetystock wellsized?Is there a transport issue?Is the distribution forecastcorrect?Arbitration issue?Is the Industrialbuffer wellsized?Production issue?Tacticalforecastissue?Efficiency lossPartial

29、vision no clear visionAction on wrong causesAAA:TOO MUCH DATA AND EVENTS TO ASSESS AND DRIVE SC PERFORMANCE20 data to analyse for each order not full filled4 millions data to analyse per week with a service level at 80%NEW FIELD OF DATA USAGE :AAA(ACTIONNABLE ADVANCED ANALYTICS)TO FOCUS USERS ON REL

30、EVANT ACTIONS FOR OPERATIONAL EXCELLENCEAAA give automatic root causes analysis and relative weight for KPI,with comprehensive data accumulation.Efficiency and lucidityTo act on what matters,globally and locally.Substainable reference for performance steering report.Comparaison ATF betweenEUC and EUN cause&Act on Qualit of forecastEUC