View
215
Download
0
Tags:
Embed Size (px)
Citation preview
““Intelligent”Intelligent”Business Intelligence SolutionsBusiness Intelligence Solutions
Being more Productive with Microsoft BI ToolsBeing more Productive with Microsoft BI Tools
Robert DennettStrategic Information Manager
AgendaAgenda Personal historical perspectivePersonal historical perspective BI and relational data managementBI and relational data management Analysis Server power intelligentlyAnalysis Server power intelligently
Handling time within OLAP cubesHandling time within OLAP cubes Merging / integrating data within OLAPMerging / integrating data within OLAP Data Mining ModelsData Mining Models
Delivering BI /OLAP intelligentlyDelivering BI /OLAP intelligently Sending OLAP “offline”Sending OLAP “offline” OLAP over the webOLAP over the web Linking with SharePoint Linking with SharePoint
Please ask Please ask questions at questions at
any time!!any time!!
My 20 Year BI JourneyMy 20 Year BI JourneyWhy things are now so exciting …Why things are now so exciting …
The Vision is BornThe Vision is Born
John Little (1971) founder of IRI Express softwareJohn Little (1971) founder of IRI Express software
“Models and Managers: the concept of a decision calculus ““Models and Managers: the concept of a decision calculus “
Elusive “Express” years – What’s a Prime?Elusive “Express” years – What’s a Prime?
““Powerplay” - to the desktop but no scale..Powerplay” - to the desktop but no scale..
““Holos” – when being “co-operative” becomes difficult!Holos” – when being “co-operative” becomes difficult!
““PC Express” – Solid, but alas an island!PC Express” – Solid, but alas an island!
The Odds and Sods ….The Odds and Sods ….
Microsoft OLAP Server Arrives (1999) Microsoft OLAP Server Arrives (1999)
Analysis ServerAnalysis ServerThe BI “Engine Room” …..The BI “Engine Room” ….. New levels of price performance - scalableNew levels of price performance - scalable
Open interfaces – back, front and middleOpen interfaces – back, front and middle
Powerful hierarchy & calculation functionsPowerful hierarchy & calculation functions
Data mining toolsData mining tools
Easy to deploy & useEasy to deploy & use
Challenge: Be “intelligent” in use. Challenge: Be “intelligent” in use.
Deliver the business benefits – at last!Deliver the business benefits – at last!
Analysis ServerAnalysis ServerMore than a data server….More than a data server….
DataData(Relational)(Relational)
DataData(MOLAP)(MOLAP)
DataData(Reports)(Reports)
LimitedView of
Analysis Server
TransactionsTransactions
Microsoft BI ToolsMicrosoft BI ToolsAdd “intelligence” at every layer ….Add “intelligence” at every layer ….
Relational DataRelational Data(SQL Server)(SQL Server)
MOLAPMOLAPROLAPROLAPHOLAPHOLAP
(Analysis Server)(Analysis Server)
Knowledge / InsightsKnowledge / Insights(SharePoint / Aladdin (SharePoint / Aladdin MS Office + More..)MS Office + More..)
Advanced AnalyticsAdvanced AnalyticsData MiningData MiningForecastingForecastingData mergingData mergingData EnrichmentData Enrichment
Intelligence NavigationIntelligence NavigationDelivery over the NetDelivery over the NetOff line OLAPOff line OLAP
Our End-to-End BI SolutionOur End-to-End BI SolutionIntelligently leverage tools to lower incremental investmentIntelligently leverage tools to lower incremental investment
Oracle
Unix
PC BasedTransaction
Other
AS/400.
DT
S L
oad &
Map D
ata
CompassSQL
WarehouseOLAP
ShareP
ointC
ompass P
ortalDocLibraries
Win2KOfficeSOE
Ftp
RPC
Win Copy
Http
Our Domain
OfflineOLAP
Rem
oteW
in NT
File S
erver
RemoteWin
OfficeSOE
Other Security Domains
Firew
all
InternetNielsen COG ABS
Core Business Transactions
EmailAgent
MarketingBusiness
Dictionary
3rd Party Data
EmailAgent
Make Relational Data IntelligentMake Relational Data IntelligentAn attribute rich relational data model adds analytical power ….An attribute rich relational data model adds analytical power ….
Go well beyond transaction dataGo well beyond transaction data Capture analytical business meta data Capture analytical business meta data
Entities & Attributes = Analysis DimensionsEntities & Attributes = Analysis Dimensions Analysis Dimensions = Business ValueAnalysis Dimensions = Business Value
Be creativeBe creative Add Supplier / Customer’s view of the worldAdd Supplier / Customer’s view of the world Add Department specific attributesAdd Department specific attributes
Future proof the data design Future proof the data design Add many-to-many tables around key entities for Add many-to-many tables around key entities for
yet “unknown” attributesyet “unknown” attributes Adhere to proper normalisationAdhere to proper normalisation
Make Analysis Server IntelligentMake Analysis Server Intelligent
Three key thingsThree key things MDX – MDX – Multidimensional Expression LanguageMultidimensional Expression Language
MDX – MDX – Multidimensional Expression LanguageMultidimensional Expression Language
MDX – MDX – Multidimensional Expression LanguageMultidimensional Expression Language
Think dimensionally!!Think dimensionally!!
Benefits of MDXBenefits of MDX Eliminate procedural programming – meet changing user needs Eliminate procedural programming – meet changing user needs
fastfast Cost off the client – the analytical power is on the serverCost off the client – the analytical power is on the server Create new options for data delivery – web / remoteCreate new options for data delivery – web / remote Very powerful functions including data miningVery powerful functions including data mining
Essential ReferencesEssential References
Using MDX to Tame “Time”Using MDX to Tame “Time”
““Time” hierarchy dimensionTime” hierarchy dimension Day / Week / Month / YearDay / Week / Month / Year
Common OLAP problem is how best to Common OLAP problem is how best to manage “Time” aggregates?manage “Time” aggregates? Year to Date - YTDYear to Date - YTD Month to Date - MTDMonth to Date - MTD Moving Annual Total – MATMoving Annual Total – MAT
Solutions ?Solutions ? Aggregate & store “time” totals as MOLAP dataAggregate & store “time” totals as MOLAP data User sum “time” members in the client.User sum “time” members in the client. Create a “Time Utility” dimension & unleash MDXCreate a “Time Utility” dimension & unleash MDX
Time “Utility” DimensionTime “Utility” DimensionUse MDX to create relative time aggregates on the fly….Use MDX to create relative time aggregates on the fly….
Customer
PeriodicityPeriodicity
Time
Year
Month
Week
Day
CurrentCurrent YTDYTD MTDMTD MATMAT
Time “Utility” DimensionTime “Utility” DimensionUse MDX to create relative time aggregates on the fly….Use MDX to create relative time aggregates on the fly….
Customer
PeriodicityPeriodicity
Time
Year
Month
Week
Day
CurrentCurrent YTDYTD MTDMTD MATMAT
Year to Dateas at each
Day/Week/Monthfor every customer
No Groupingof time
MDX Syntax SamplesMDX Syntax Samples
Create a new “Periodicity” calculated memberCreate a new “Periodicity” calculated member YTD – sum from first day of Year to current timeYTD – sum from first day of Year to current time
MAT – sum from this day a year ago (plus 1day) to MAT – sum from this day a year ago (plus 1day) to current timecurrent time
SUM ( PeriodsToDate([Time].[Year], [Time].CurrentMember) )SUM ( PeriodsToDate([Time].[Year], [Time].CurrentMember) )
SUM( { ParallelPeriod( [Time].[Year], 1, [Time].CurrentMember).NextMember SUM( { ParallelPeriod( [Time].[Year], 1, [Time].CurrentMember).NextMember : [Time].CurrentMember } ) : [Time].CurrentMember } )
Analysis Server Data Mining Tools Analysis Server Data Mining Tools Mine Data Attributes to Extract Hidden KnowledgeMine Data Attributes to Extract Hidden Knowledge
Out of the box functionality
Which input(s) best
“explain” or predict changes
is some key output measure.
Tightly integrated feature
End users can directly tap
benefits of data mining using
everyday tools like Excel
Sales
Sample Data Mining Output Sample Data Mining Output Top 2 factors …Top 2 factors …
SalesSales
Sample Data Mining Output Sample Data Mining Output Top 3 factors …Top 3 factors …
SalesSales
Use MDX to Take Analysis “Offline”Use MDX to Take Analysis “Offline”& onto the road& onto the road Use MDX to create “cube files” which are Use MDX to create “cube files” which are
then independent of Analysis Serverthen independent of Analysis Server Powerful method of delivering remote BIPowerful method of delivering remote BI
Laptop users – sales repsLaptop users – sales reps Share BI with customers & suppliersShare BI with customers & suppliers Where web browsing is NOT possibleWhere web browsing is NOT possible Use same client tools as server cube.Use same client tools as server cube.
Local cubes have limitations. Plan ahead to Local cubes have limitations. Plan ahead to avoid theseavoid these Size – calculated members become real dataSize – calculated members become real data Security – there is none!Security – there is none! Default members and roles can cause a few Default members and roles can cause a few
problemsproblems Create Cube syntax is somewhat complexCreate Cube syntax is somewhat complex
Creating “Offline” OLAP CubesCreating “Offline” OLAP CubesHarness powerful low cost 3Harness powerful low cost 3rdrd party tools party tools
Local Cube TaskSDG Computing, Inc. http://www.localcubetask.com/index.htm
SQL Server DTS Add-in
“LocalCube Task”
Complete “shrink wrap”
solution :-
•Defining local cubes
•Production updates in DTS
•Managing Cube Distribution
Compression
File copy
•Build & pack OWC web
pages along with data
Local Cube TaskLocal Cube TaskHighly functional tool at bargain basement prices!!Highly functional tool at bargain basement prices!!
Take Analysis onto the WebTake Analysis onto the Web
Analysis Server / IIS can be linked Analysis Server / IIS can be linked through ASP through ASP
Very powerful method of delivering BI Very powerful method of delivering BI using only a web browserusing only a web browser Zero client - HTMLZero client - HTML Empower wireless laptop / PDA usersEmpower wireless laptop / PDA users Customers, suppliers, sales reps….Customers, suppliers, sales reps….
OLAP over the web has few limitationsOLAP over the web has few limitations Security – authentication firewall issues Security – authentication firewall issues User expectations for fancy graphicsUser expectations for fancy graphics
Take OLAP to the WebTake OLAP to the WebAgain, harness powerful low cost 3Again, harness powerful low cost 3rdrd party tools party tools
IIS ASP EngineIIS ASP Engine Takes MDX Takes MDX
queries, renders queries, renders HTML outputHTML output
Includes starter Includes starter templatestemplates
SQL Report SQL Report RepositoryRepository
Write you own Write you own ASP pagesASP pages
Strategic Data Management’s “Aladdin”http://aladdin.sdm.com.au/
OLAP Over the WebOLAP Over the Web
Delivering Integrated BIDelivering Integrated BIUsing a SharePoint Portal as “BI umbrella”Using a SharePoint Portal as “BI umbrella”
All data cannot be integrated into one cubeAll data cannot be integrated into one cube BI “using experience” can be integrated into BI “using experience” can be integrated into
single using environmentsingle using environment OLAP reports (hard data)OLAP reports (hard data) Business plans, research reports (soft data)Business plans, research reports (soft data) Blended – business plans with dynamic OLAP Blended – business plans with dynamic OLAP
contentcontent
Avoid user BI tool clutter Avoid user BI tool clutter The power is on the server not the client!The power is on the server not the client! Reduce BI costs significantlyReduce BI costs significantly A case for no more than two toolsA case for no more than two tools
MS Office / Excel MS Office / Excel 1 Other1 Other
Linking Analysis Server & SharePointLinking Analysis Server & SharePointAdd Intelligence to Add Intelligence to SharePoint SubscriptionsSharePoint Subscriptions
Use SQL DTS Use SQL DTS Package Package Update OLAP cubeUpdate OLAP cube
Refresh pivot tableRefresh pivot table
““Check In” to Check In” to SharePointSharePoint
SharePoint SharePoint “subscribers” receive “subscribers” receive notification email notification email with URL to new datawith URL to new data
Final Thoughts - Dr Codd (1994)Final Thoughts - Dr Codd (1994)White paper in which he developed the term “OLAP”…White paper in which he developed the term “OLAP”…
““Attempting to force one technology or Attempting to force one technology or tool to satisfy a particular need for tool to satisfy a particular need for which another tool is more effective which another tool is more effective and efficient is like attempting to drive a and efficient is like attempting to drive a screw into a wall with a hammer when a screw into a wall with a hammer when a screwdriver is at hand:screwdriver is at hand:
…. the screw may eventually enter the …. the screw may eventually enter the wall but at what cost?” wall but at what cost?”
Thank you Dr Codd….Thank you Dr Codd….