SESAM Business Intelligence og Automation June 2012 Gennemgang af forskellige Business Intelligence projekter og løsninger, med fokus på udnyttelse af flere datakilder i en produktionsvirksomhed.
Poul W. Nielsen Senior Consultant, NNE Pharmaplan A/S, 2002-Present Production Manager, Toms Fabrikker A/S IT Consultant, NCN Systems A/S Service Manager, A/S Tetra Pak IT Consultant, Own Company M.Sc. Engineering 1989 Poul Waagner Nielsen Senior Consultant, Automation & IT Consulting MSc Eng. NNE Pharmaplan Hallas Allé 1, DK-4400 Kalundborg, Denmark pwni@nnepharmaplan.com Telephone: +45 4444 7777, Direct: +45 3075 3993, Mobile: +45 3075 3993
Agenda 1. Introduktion 1. Cases overview 2. Basics in Business Intelligence 3. Cases 1. KPI/Dashboard platform w. multiple sources 2. Lightweight OEE system w. Single source 3. Scrap KPI system with challenging data dimensioning
BI - Elements Get Data Prime challenge: Safe/Datadriven normalisation (keys/dimension entries) is the prime challenge for succes/durability. Dimensioning,normalisation Get Data Out Reporting, vizualisation is simple work if normalisation is in place. Source: tdwi.org & MS BI samples 4
Cases 1. KPI/Dashboard platform w. multiple sources 2. Scrap KPI system with challenging data dimensioning 3. Lightweight OEE system w. single source Common for all 3 cases: Hybrid Business Intelligence Application & Business Inteligence in one-solution with manuel data enrichment in integrated web frontend. Developed on-site togheter with customer in a Scrum like project setup. Gives max. flexibility and iterative development/user expierience process loops. Based on SQL Server/Reporting Services/asp.NET User configurable mapping of source data to key s Flexible, safe, datadriven and simplifies BI project. No use of traditional Cube s Works diretcly on relational data emulating cubes via keys and dimension master tables. Reduces complexity, development cost, load cycles Enables application like user expierience. PM2 (Novo Nordisk) KPI platform for laboartories TSM (Novo Nordisk) Production Scrap KPI system DBS (Axellus) Lightweight OEE System 5 Pharmaceutical laboratories - Automation and IT services
Case 1 PM2 Performance Management system Global novo Nordisk Lab-KPI system Active for > 3 years > 1000 active Users Fully manageable by super users. Dozens of KPI s/dashboards Hybrid KPI/Application system Sporting advanced distributed data enrichment Platform All on one server SQL Server 2008R2 Reporting Services IIS/Asp.NET application 6
Case 1 PM2 Landscape ETL - Automatic loads CORP net Web Dashboards & reporting Data source ERP SAP H/R LIMS Data transfer transfer & wash SQL Server RDBMS Warehouse Operational Data Storage Cube emulation Reporting Serv. Asp Frontend Master Data & Enrichment Microsoft Reporting services sports: MES Automated ready output to ext. Systems Easy user access control Easy mail subscription (get reports via email) Drill down facilities Full export to Excel, PDF, text e.g. And more through simple web interface HPLC Other datasources can be added Time registrations in dedicated easy to use web application 7
Case 1 - PM2 Performance Management system All-in-one frontend Administrator menu - Masterdata maintenance - Dataloads - User admin. All in one-frontend Common user experience Distributed MD setup. Minimum central admin Easy expandable User menu KPI Reports Data enrichment Dashboards 8
Case 1- PM2 Performance Management system Examples on setup in integrated web application User enabled - Master data dimension setup one entry per unique key. User driven keymapping. Distributed to 100 different departments. User enabled Goal Settings Custom made KPI s available for user to make. 9
Case 1 - PM2 Performance Management system Data enrichment Data enrichment on incoming LIMS data. Enables distributed maintenance of supplemental data in the vast laboratory organisation. Registration of data on custom made KPI s. Enables a varity of custom KPI s defined by and for the users/departments suiting their own permanent or ad/hoc needs. Unique integrated data enrichment facilities Enables post-categorizing of automatic loaded data distributed to right users. 10
Case 1 - PM2 Performance Management system Multiple dashboards covering prime laboratory focus HPLC performance measuring Measures combines LIMS source data with user categorized data via PM2 application. Productivity based on SAP H/R Hours and Produced items Full drill down Contributor s LIMS, Custom KPI s a.o. 11
Case 2 Scrap KPI system with challenging data dimensioning Admin center. Load, Parse rule setup Goal settings Support data for key evaluation Master key data Reports (available to users) Novo Nordisk Production Scrap-KPI system Fully manageable by super users. Data Wash/Key-Normalisation user-configurable Hybrid KPI/Application system Distributed data enrichment Replaces Excel solution Platform All on one server SQL Server 2008R2 Reporting Services IIS/Asp.NET application 12
Case 2 Scrap KPI system with challenging data dimensioning
Case 2 Scrap KPI system User driven load, wash & dimensioning Status on MasterData Key alignment. User center for Wash&Normalization Load of multiple BW Sources Control Key Normalization Store data to historic Key / Dimensions Measures Ready to dice/report on 14
Case 2 Scrap KPI system Advanced user controlled data dimensioning Parsesequence Each Key in gets data from it s value statement that is parsed during Stage-#2 Statement deriving MaterialType to place in key. Result must match corresp. Master key table. Challenging normalization Source data from 2 different ERP systems given via BW Very complex rules for key/dimension evaluation. Interdependent keys Untraditional solution Open Parse statements make it possible In a flexible way to normalize data in dimensions not available in any sourcedata yet still controlled by user. To align 20 keys/measures in 5 different sources with each it s special parse-statement. Deliver all data into one single historic table whereupon reporting takes place. Enables the user to adapt changes over time in organisation, rules that is not reflected by the datasources. 15 2012-06-11
Case 2 Scrap KPI system - Reporting examples Dice and filter on all KeyDimensions created via parsing statements. Full drill down 16
Case 3 Lightweight OEE system / Axellus A/S Administrator menu - Masterdata maintenance - User admin. All in one-frontend Common user experience Distributed MD setup. Minimum central administration Easy expandable User menu KPI Reports Data enrichment Dashboards Axellus A/S DBS data baseret stop system System. Active for > 2 years Fully managable by super users. Hybrid KPI/Application system Distributed data enrichment Platform All on one server SQL Server 2008R2 Reporting Services IIS/Asp.NET application 17
Case 3 OEE System F.x. Intouch DB Læser tællere fra linier hvert minut. (hvis muligt) Data transfer & wash SQL Server RDBMS Warehouse Operational Data Storage Cube emulation Reporting Serv. User inteface. Local net Rapportering tilrådighed for alle på netværk Rapporter 4-5 rapporter efter ønsker/muligt f.x.: Top10 stop,oppetid, trend, detaljedata. Fremsøgning og aggrering dimensioneret på: Produkt/Produktgruppe Linie/Produktionsafsnit År/Kvartal/Måned/Uge/Dag Stopkode/Stopkodegruppe Microsoft Reporting services sports: Produktionsregistrering: Linie,varenummer, stk/minut (*), start, stop Stopregistrering: Stopkode, varighed, bemærkning, tid for indtastning. (er der på linie aut. Tæller i intouch hentes stop her og bruger angiver blot stopkode) (*) angives manuelt ved hver produktion. Stk/minut findes ikke i masterdata. Master data setup: Produktionslinier,Varenumre, Stopkoder Øvrig setup af administrator direkte i rå database. (produktionslinier kan tilføjes vedr start eller senere hvis flere tilkommer). Easy user access control Easy mail subscription (get reports via email) Drill down facilities Full export to Excel, PDF, text e.g. Saving og charts in file downto once a minute enabling displaing on exernal pc/screens And more through simple web interface
Case 3 Lightweight OEE system Setup & Data enrichment / Report examples. Master data setup. Lines, Users, Goals. User driven keymapping of lines to MES source (counter). OEE & Oppetid Waterfall Data enrichment. Add stop reason codes. Distributed to users direct on lines. Online trending OEE,Stop. Full Dimension Dice&Select 19