23
SharePoint 2010 Business Intelligence Module 6: Analysis Services

SharePoint 2010 Business Intelligence Module 6: Analysis Services

Embed Size (px)

Citation preview

Page 1: SharePoint 2010 Business Intelligence Module 6: Analysis Services

SharePoint 2010 Business Intelligence

Module 6: Analysis Services

Page 2: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Overview

Analysis Services

Page 3: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Lesson: Analysis Services

Introduction

ETL

OLAP Terms

Storage Modes

Queries

Tools

Mining Models

Page 4: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Introduction

Analysis Services provides access to large data sets

Running SQL queries against a 100 million row table just doesn’t work

When your data sets get large, you need a better way of handling the load

Online analytical processing (OLAP) provides all your answers with speed!

Analysis services is an OLAP implementation

Page 5: SharePoint 2010 Business Intelligence Module 6: Analysis Services

ETL

The Extract, Transform and Load (ETL) process is vital to the OLAP results

If you input junk, you get junk back out! All data that enters the OLAP database must be valid or

your results could be exponentially wrong!Make sure that all possible means are employed when ensuring only valid data is entered into the system and that it is entered only once!

Common issue is data inserted more than once which causes invalid results

Page 6: SharePoint 2010 Business Intelligence Module 6: Analysis Services

OLAP Terms

Data Source

A source or destination of data Fact table

A table that contains numerical information with keys that map to the dimension keysMeasure

A single numerical value in a fact tableDimension

A set of labels/attributes that describe the measures in a fact tableCube

A set of aggregations of all the dimensions and facts complied together to produce valuable information

Perspective

A subset of dimensions and measures specific to some group of usersData Modeling

Using statistical analysis to determine patterns in large sets of data

Page 7: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Tools

Visual Studio / BI Workbench

Several project templates to facilitate the creation of cubes, dimensions and work with data sources/views

IntelliCube

A heuristic analysis tool for automatic generation of a cube based on its data and relationships

External Viewer forOutliersCandidate keysValue distributionsPatterns

Page 8: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Business Intelligence Development Studio

BI Studio is just Visual Studio with project templates installed

Project templates provide item templates and wizards

New 2008 Wizards are much easier to use and more powerful

Context sensitive functionality keeps toolbars and menus trimmed to what elements you need based on what you are looking at

Because it is Visual Studio it is fully customizable

3rd party add-ons can make it even more powerful

Page 9: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Designing an Analysis Services Database

BI Studio will be used to create new AS Databases

Steps include:

Define data source Define data view Create a new cube Define fact tables Define dimensions (data and time) Define measures Aggregate/Run the cube

Page 10: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Key Performance Indicators (KPI)

KPIs are used to show very simply whether a target is being accomplished

Components include the Goal, Value, Status, and Trend Example: Sales and Quality targets

Analysis Services allows you to build MDX expressions off of Cube data to build KPIs

KPI values can be queried from client applications Allows for visual display of meaningful data

Page 11: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Actions

Actions

Allows client application users to be able to interact with what the data means

Example: Browse to a customer or product via URLTypes of Actions:

CommandLine, DataSet, Drillthrough, Html, Custom, Report, URL

Define and assign to objects in the Cube

Page 12: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Perspectives

Perspectives work similar to views in a relational database

Shows different users the data they need to see for a particular role they may be in

Used to reduce complexity of cube data

Not meant to be security mechanism for data

Page 13: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Storage Modes

Data in an Analysis Services database is stored differently than a relational database

Optimized storage provides the OLAP query performancePartition

MOLAP – multidimensional OLAP (fact data and aggregations are stored in special format)

ROLAP – Relational OLAP (fact data and aggregations remain in relational database)

HOLAP – Hybrid OLAP (fact data is relational, aggregations are stored in special format)

Dimension (dimension attributes only)

MOLAP – stored in special format ROLAP – stay in relational format

Page 14: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Querying Cubes

SQL is not used in OLAP databases

MDX (multidimensional queries) is used for querying cubes

DMX (Data Mining queries) is used for querying data mining models

Page 15: SharePoint 2010 Business Intelligence Module 6: Analysis Services

MDX Queries

Multidimensional Expressions (MDX) are used to query multidimensional data

Some common terms are:

Cell – the space at an intersection of a measure and attribute

Tuple – a unique cell based on a set of attribute members Set – an order set of tuples with same dimensionality

Page 16: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Calculated Members and Named Sets

Calculated Members are used when you need to determine something at query time

Can be query or session scoped Value are only stored in memory not on disk

Named Sets are basically predefined MDX queries that can be reused in other queries

Used to group dimension members

Page 17: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Analyzing Data with Data Mining Algorithms

Out of the box, Analysis services provides five algorithms:

Classification (Decision Tree)Predict one or more discrete variables, based on the other attributes in the dataset

Regression (Time Series)Predict one or more continuous variables, such as profit or loss, based on other attributes in the dataset

Segmentation (Clustering)Divide data into groups, or clusters, of items that have similar properties.

Association (Association)Find correlations between different attributes in a dataset

Sequence analysis (Sequence Clustering)Summarize frequent sequences or episodes in data, such as a Web path flow

Page 18: SharePoint 2010 Business Intelligence Module 6: Analysis Services

What could go wrong?

Cubes really are simple things to build and utilize

As simple as they are, it is easy to create a cube that has the wrong data

Always validate that the data that is displayed in the Cube is valid and accurate

Never hurts to be overly aggressive when testing Cube data

Page 19: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Lab 1: Analysis Services

Explore Analysis Services

Page 20: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Lab 2: Building a Cube

Explore Cubes and Dimensions

Page 21: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Lab 3: Data Mining Algorithms

Explore Data Mining Models

Page 22: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Review

Your instructor will ask a series of questions on this module

Page 23: SharePoint 2010 Business Intelligence Module 6: Analysis Services

Summary

Extra Large databases are not easily queried for data

Analysis Services is an OLAP tool to manage large databases

Ensure that your ETL process is accurate

Data Mining Algorithms can help you find patterns you didn’t know about before