1 Chapter 1: Overview 1.1 Overview of Business Analytics 1.2 Software Used in This Course 1.3 Data Management 1.4 Recommended Reading.

1

Chapter 1: Overview

1.1 Overview of Business Analytics

1.2 Software Used in This Course

1.3 Data Management

1.4 Recommended Reading

2

Chapter 1: Overview



1.3 Data Management


3

Objectives Define analytics and data mining. Explain the proliferation of data and how this impacts

the need for good analytics. Identify some of the key challenges of data mining. Name some applications where analytics are helpful. Name some applications where analytics are not

helpful. Explain some of the common pitfalls of analytical

practice.

4

Analytics“The extensive use of data, statistical and quantitative analysis, explanatory and predictive models, and fact-based management to drive decisions and actions.”

Davenport and Harris (2007)

Competing on Analytics:

The New Science of Winning

5

Co

mp

etit

ive

Ad

van

tag

e

Basic Reporting What happened?

Ad Hoc Reporting How many, how often, where?

Dynamic Reporting Where exactly are the problems?

Reporting with Early Warning What actions are needed?

Basic Statistical Analysis Why is this happening?

Forecasting What if these trends continue?

Predictive Modeling What will happen next?

Decision Optimization What is the best decision?

Data Information Intelligence

Advanced Analytics

Basic Analytics

Reporting

Decision Support Decision Guidance

Achieving Success with Analytics

6

elec

troni

c po

int-o

f-sal

e da

ta

hosp

ital p

atie

nt re

gist

ries

cata

log

orde

rs

ban

k tra

nsac

tions

rem

ote

sens

ing

imag

es

tax

retu

rns

airli

ne re

serv

atio

ns

cre

dit c

ard

char

ges

stoc

k tra

des

OLT

P t

elep

hone

cal

ls

Data Deluge

W

eb c

omm

ents

7

Three Consequences of the Data Deluge1. Every problem will generate data eventually.

2. Every company will need analytics eventually.

3. Everyone will need analytics eventually.

...

8


Proactively defining a data collection protocol will result in more useful information, leading to more useful analytics.

2. Every company will need analytics eventually.


...

9



2. Every company will need analytics eventually.Proactively analytical companies will compete more effectively.


...

10



2. Every company will need analytics eventually.Proactively analytical companies will compete more effectively.

3. Everyone will need analytics eventually.Proactively analytical people will be more marketable and more successful in their work.

11

The Business Analytics ChallengeGetting anything useful out of tons and tons of data

12

elec

troni

c po

int-o

f-sal

e da

ta

hosp

ital p

atie

nt re

gist

ries

stoc

k tra

des

OLT

P t

elep

hone

cal

ls

Hope for the Data Deluge

= actionable knowledge

+ analytical tools

W

eb c

omm

ents

airli

ne re

serv

atio

ns

cre

dit c

ard

char

ges

cata

log

orde

rs

ban

k tra

nsac

tions

rem

ote

sens

ing

imag

es

tax

retu

rns

13

Changes in the Analytical Landscape

Analytical Modelers Management

Historically…

Historically, analytics have typically been handled in the “back office,” and information was shared only by a few individuals.

Models

14

Changes in the Analytical LandscapeHistorical Changes Executive Dashboarding – Static reports on business

processes Total Quality Management (TQM) – Customer focused Six Sigma – Voice of the process, Voice of the

customer Customer Relationship Management (CRM) – The

right offer to the right person at the right time Forecasting and Predicting – 360-degree customer

view

15

Changes in the Analytical LandscapeRelational Databases

Enterprise Resource Planning (ERP) Systems

Point of Sale (POS) Systems

Data Warehousing

Decision Support Systems Reporting and Ad Hoc Queries Online Analytical Processing (OLAP)

Performance Management Systems Executive Information Systems (EIS) Balanced Scorecard Dashboard

Business Intelligence

16

CRM EvolutionTotal Quality Management (TQM)

Product Centric Quality: Six Sigma Total Customer Satisfaction Mass Marketing

One-to-One Marketing

Customer Relationship Wallet Share of Customer Customer Retention

Customer Relationship Management (CRM)

Customer Centric Strategy Process Technology

17

Changes in the Analytical Landscape

Analytical Modelers

CustomerService

Retail

Logistics

Promotions

OPERATIONS TARGET

Customers

Stockholders

Suppliers

Employees

Now…

Now analytics are being pushed out to the “front office” and are directly impacting company performance. There are clear, tangible benefits that management will track. Data mining is a critical part of business analytics.

Proliferation of Models

18

Idiosyncrasies of Business Analytics1. The Data

– Massive, operational, and opportunistic

2. The Users and Sponsors– Business decision support

3. The Methodology– Computer-intensive adhockery– Multidisciplinary lineage

19

The DataExperimental Opportunistic

Purpose Research Operational

Value Scientific Commercial

Generation Actively Passivelycontrolled observed

Size Small Massive

Hygiene Clean Dirty

State Static Dynamic

20

The Data: Disparate Business Units

Marketing Invoicing Risk

Acquisitions Operations Sales

21

Opportunistic Data Operational data is typically not collected with data

analysis in mind. Multiple business units produce a silo-based data

system. This makes business analytics different from

experimental statistics and especially challenging.

22

The Methodology: What We Learned Not to DoPrediction is more important than inference. Metrics are used “because they work,” not based on

theory. p-values are rough guides rather than firm decision

cutoffs. Interpretation of a model might be irrelevant. The preliminary value of a model is determined by its

ability to predict a holdout sample. Long-term value of a model is determined by its ability

to continue to perform well on new data over time. Models are retired as customer behavior shifts, market

trends emerge, and so on.

23

Using Analytics IntelligentlyIntelligent use of analytics results in the following: Better understanding of how technological, economic,

and marketplace shifts affect business performance Ability to consistently and reliably distinguish

between effective and ineffective interventions Efficient use of assets, reduced waste in supplies, and

better management of time and resources Risk-reduction via measurable outcomes and

reproducible findings Early detection of market trends hidden in massive

data Continuous improvement in decision making over time

24

Simple ReportingExamples: OLAP, RFM, QC, descriptive statistics, extrapolation

Answer questions such as

Where are my targets now?

Where were my targets last week?

Is the current process behaving like normal?

What’s likely to happen tomorrow?

25

Proactive Analytical Investigation Examples: inferential statistics, experimentation, empirical validation, forecasting, optimization

Answer questions such as

What does a change in the market mean for my targets?

What do other factors tell me about what I can expect from my target?

What is the best combination of factors to give me the most efficient use of resources and maximum profitability?

What is the highest price the market will tolerate?

What will happen in six months if I do nothing? What if I implement an alternative strategy?

26

Data StalemateMany companies have data that they do not use or that is used by third parties. These third parties might even resell the data and any derived metrics back to the original company!

Example: retail grocery POS card

27

Every Little Bit…Taking an analytical approach to only a few key business problems with reliable metrics tangible benefit.

The benefits and savings derived from early analytical successes managerial support for further analytical efforts.

Everyone has data.

Analytics can connect data to smart decisions.

Proactively analytical companies outpace competition.

28

Areas Where Analytics Are Often UsedNew customer acquisition

Customer loyalty

Cross-sell / Up-sell

Pricing tolerance

Supply optimization

Staffing optimization

Financial forecasting

Product placement

Churn

Insurance rate setting

Fraud detection

…

Which residents in a ZIP code should receive a coupon in the mail for a new store location?

29


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

What advertising strategy best elicits positive sentiment toward the brand?

30


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

What is the best next product for this customer?

31


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

What is the highest price that the market will bear without substantial loss of demand?

32


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

How many 60-inch HDTVs should be in stock? (Too many is expensive; too few is lost revenue.)

33


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

What are the best times and best days to have technical experts on the showroom floor?

34


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

What revenue increase can be expected after the Mother’s Day sale?

35


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

Will oatmeal sell better near granola bars or near baby food?

36


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

Which customers are most likely to switch to a different wireless provider in the next six months?

37


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…

How likely is it that this individual will have a claim?

38


Customer loyalty


Pricing tolerance

Supply optimization



Product placement

Churn


Fraud detection

…How can I identify a fraudulent purchase?

39

When Analytics Are Not HelpfulSnap decisions required

Novel approach (no previous data possible)

Most salient factors are rare (making decisions to work around unlikely obstacles or miracles)

Expert analysis suggests a particular path

Metrics are inappropriate

Naïve implementation of analytics

Confirming what you already know

Deciding when to run from danger

40








Predicting the adoption of a new technology

41








Planning contingencies for employees winning the lottery

42








The seasoned art critic can recognize a fake

43








Predicting athletes’ salaries or quantifying love

44







Confirming what you already knowOnly looking at one variable at a time

45

When Analytics Are Not HelpfulSnap-decisions required






Confirming what you already know Ignoring variables that might be important

46

Naïve AnalyticsMany companies implementing analytical programs such as Six Sigma demonstrate tremendous success.

However, it is important to use analytics in a meaningful way.

For example: It might not be possible to establish a Six Sigma

process with low production volume. Producer-centric metrics might not give useful

information about customer satisfaction, and the Six Sigma process might still fail to meet customer specifications.

Simplistic reporting on massive data might hide complex patterns and is generally unsuccessful.

47

The Fallacy of Univariate Thinking

What is the most important cause of churn?

Prob(churn)

InternationalUsage

DaytimeUsage

48

Expectations Leading the AnalysisEven sophisticated analytics are not immune to personal bias such as selectively fitting models with variables because they

place someone’s opinion or agenda in a positive light ignoring information that might disprove a hypothesis.

Personal bias in model fitting, whether intentional or otherwise, can diminish the usefulness of your analytical efforts.

49

Trustworthy AnalyticsLet the data guide your conclusions.

Ask the following questions: Are my assumptions about the causes of my data

patterns warranted? Should I be trying something different?

Assign a cynic to the analytical team whose purpose is to question the assumptions. What would my critic say is the flaw with my analysis? Investigate the data in such a way that a critic’s

concerns can be ruled out.

50

Chapter 1: Overview



1.3 Data Management


51

Objectives Identify and describe several software tools for

business analytics, including – SAS Enterprise Guide– SAS Enterprise Miner– SAS Forecast Studio.

Describe several of the key features of these programs.

52

Tasks for Data Analysis Access data sources Combine tables Transform variables Explore and describe data Visualize patterns in the data Analyze and model data Validate models Score Update models

53

The Tools of Data Analysis

...

54


...

55


56

Introduction to the SAS SystemThe SAS System is driven by SAS programs. SAS programs are composed of commands in the form

of DATA steps and PROC (or procedure) steps. The SAS System features point-and-click interfaces to

SAS that write programs and perform additional functions automatically. Three interfaces featured in this course are– SAS Enterprise Guide– SAS Enterprise Miner– SAS Forecast Studio.

57

Introduction to SAS Enterprise Guide

58

Introduction to SAS Enterprise GuideSAS Enterprise Guide provides a point-and-click interface for managing data and generating reports.

59

Introduction to SAS Enterprise GuideBehind the scenes, SAS Enterprise Guide generates SAS programs that are submitted to SAS, and results are returned to SAS Enterprise Guide.

60

SAS Enterprise Guide InterfaceSAS Enterprise Guide also includes a full programming interface that can be used to write, edit, and submit SAS programs.

61

SAS Enterprise Guide Interface: The ProjectA project serves asa collection of data sources SAS programs

and logs results from

tasks and queries informational notes

for documentation.

You can control the contents, sequencing, and updating of a project.

62

Building a SAS Enterprise Guide ProjectTo begin work with SAS Enterprise Guide, you

1. create a new project

2. add data to the project

3. run tasks against the data.

In addition, you can

4. customize results

5. automate the process.

63

Behind the ScenesAs you build tasks, SAS Enterprise Guide generates SAS programs.

64

The SAS Help FacilityUse the SAS Enterprise Guide Help facility for additional direction about SAS Enterprise Guide. Additional help is available online.

http://support.sas.com/documentation/onlinedoc/guide

http://support.sas.com/documentation/onlinedoc/guide

65

SAS Enterprise Miner

66

SAS Enterprise Miner: Interface Tour

Menu bar and shortcut buttons

67


Project panel

68


Properties panel

69


Help panel

70


Diagram workspace

71


Process flow

72


Node

73


SEMMA tools palette

74

SAS Enterprise Miner Analytic Strengths

Predictive Modeling

Pattern Discovery

75

SAS Software for Forecasting Base SAS

– DATA step programming SAS/STAT

– Ordinary least squares regression– Regression with correlated errors

SAS/ETS– Univariate and multivariate time series forecasting– Dynamic regression with correlated errors– Econometric modeling– Spectral analysis– Time series cross-sectional modeling

Forecast Server– High performance large-scale forecasting

...

76

SAS Software for Forecasting Base SAS

– DATA step programming SAS/STAT

– Ordinary least squares regression– Regression with correlated errors

SAS/ETS– Univariate and multivariate time series forecasting– Dynamic regression with correlated errors– Econometric modeling– Spectral analysis– Time series cross-sectional modeling

Forecast Server– High-performance large-scale forecasting

77

Environments for Forecasting Using SAS The SAS windowing environment The Time Series Forecasting System SAS Enterprise Guide SAS Forecast Studio

...

78

Environments for Forecasting Using SAS The SAS windowing environment The Time Series Forecasting System SAS Enterprise Guide SAS Forecast Studio

79

SAS Forecast Studio: Interface Tour

80

SAS Forecast Studio: Interface Tour Menu bar and shortcut buttons

81


Active series

Overview panel

82


Four view tabs

83


The Forecasting view

84


The Modeling view

85


The Scenario Analysis view

86


The Series view

88

Large-Scale Forecasting Scenario

Time Series Data

80% can be forecast automatically.

10% require extra effort.

10% cannot be forecast accurately.

89

IDEA EXCHANGE Identify several business problems that you could address

with analytics. You might think about case studies you are using in other classes, projects you work on at your job, examples from the media, or others.

Describe the goal, whether the variables can be measured, how the data could be obtained, and what types of specific questions you would like to address with analytics.

For now, entertain several examples. Later in the course, you might consider how to collect or download) the relevant data, perform a data analysis, and report your findings to others in an organization.

What kind of data do you think you would need? How could you obtain it?

What software tools do you think you are likely to use?

90

Chapter 1: Overview



1.3 Data Management


91

Objectives Explain the concept of data integration. Describe SAS Enterprise Guide and how it fits in to

data integration and management for business analytics.

92

Data Management Business AnalyticsData management brings together data components that can exist on multiple machines, from different software vendors, throughout the organization.

Data management is the foundation for business analytics. Without correctly consolidated data, those working in the analytics, reporting, and solutions areas might not be working with the most current, accurate data.

Advanced Analytics

Basic Analytics

Reporting

93

Managing Data for Business Analytics Business analytics requires data management

activities such as data access, movement, transformation, aggregation, and augmentation.

These tasks can involve many different types of data (for example, simple flat files, files with comma-separated values, Microsoft Excel files, SAS tables, and Oracle tables).

The data likely combines individual transactions, customer summaries, product summaries, and/or other levels of data granularity.

94

Planning from the Top Down

What data will help you answer these questions?

What mission-critical questions must be answered?

What data do you havethat will help you buildthe needed data?

...

95

Implementing from the Bottom Up

Define Target Data

Identify Source Data

Create Reports

...

96

Collaboration Is Key to Business Analytics

Business Expert IT Expert Analytical Expert

97

Data Marts: Tying Questions to DataIn a very basic form, data marts are implemented at organizations because there are questions that must be answered.

Data is typically collected in daily operations but might not be organized in a way that answers the questions.

An IT professional can use the questions and the data collected from daily operations to construct the tables for a data warehouse or data mart.

98

Building a Data MartFoundation of a Data Mart

Create target tables.

Identify target tables.

Identify source tables.

Building the foundation of the data mart consists of the three basic steps listed above.

99

Orion Star Sports & Outdoors

Orion Star Sports & Outdoors is a fictitious global sports and outdoors retailer with traditional stores, an online store, and a large catalog business.

The corporate headquarters is located in the United States with offices and stores in many countries throughout the world.

Orion Star has about 1,000 employees and 90,000 customers. It processes approximately 150,000 orders annually and purchases products from 64 suppliers.

100

Orion Star DataAs is the case with most organizations, Orion Star has a large amount of data about its customers, suppliers, products, and employees. Much of this information is stored in transactional systems in various formats.

Using SAS Enterprise Guide, this transactional information will be extracted, transformed, and loaded into a data mart for the Marketing Department.

You continue to work with this data set for some basic exploratory analysis and reporting.

101

Orion Star Source Data Normalized ModelThe Orion Star source data is organized as a normalized model.

102

A Target Star SchemaOne goal of creating a data mart is to produce, from the source data, a dimensional data model that is a star schema.

Fact TableProduct

Dimension

TimeDimension

OrganizationDimension

CustomerDimension

103

Orion Star Target Star SchemaThe analyst can produce, from the Orion Star source data, a dimensional data model that is a star schema.

OrderFactFact Table

ProductDimension

TimeDimension

OrganizationDimension

CustomerDimension

104

OrderFact Table DiagramThe OrderFact table process diagram was sketched to be this.

The source tables, target table, andjob metadata objects were already created (used for the change management demonstration), andthe target table loaded.

Customer Orders

OrderFact

Order_Item

Computed Columns

NONE

105

Product Dimension Table DiagramThe ProdDim table (representing the product dimension) process diagram was sketched to be this.

The target table and source table metadata objects must be created, as well as the job that loads the ProdDim table.

Product_List Supplier

ProdDim

Product_CategoryProduct_GroupProduct_Line

Computed Columns

106

Customer Dimension Table DiagramThe CustDim table (representing the customer dimension) process diagram was sketched to be this.

The target table and source table metadata objects must be created, as well as the job that loads the CustDim table.

Customer Customer_Type

CustDim

Customer_AgeCustomer_Age_Group

Computed Columns

107

Organization Dimension Table Diagram

Organization Staff

OrgDim

CompanyDepartment

GroupSection

The OrgDim table (representing the organization dimension) process diagram was sketched to be this.

The target table and source table metadata objects must be created, as well as the job that loads the OrgDim table.

Computed Columns

108

Time Dimension Table Diagram

TimeDim

All Columns!!(User-written code

generates theTimeDim table.)

The TimeDim table (representing the time dimension) process diagram was sketched to be this.

The target table metadata object must be created, as well as the job that loads the TimeDim table.

Computed Columns

109

Advantages of Data Marts There is one version of the truth. Downstream tables are updated as source data is

updated, so analyses are always based on the latest information.

The problem of a proliferation of spreadsheets is avoided.

Information is clearly identified by standardized variable names and data types.

Multiple users can access the same data.

110

SAS Enterprise Guide OverviewSAS Enterprise Guide can be used for data management, as well as a wide variety of other tasks: data exploration querying and reporting graphical analysis statistical analysis scoring

111

Example: Orion Star Data ManagementThe head of marketing wants to know which products are most popular in different regions. Should marketing efforts be applied to specific regions for certain products?

Create part of an analytical data mart by combining information from three tables: transactional order data, customer records, and product information.

The data for this example can be found in two types of file: SAS data sets Microsoft Excel workbook

112

Input FilesOrderinformation.sas7bdat

Productinformation.sas7bdat

Customerinformation.xls

113

Final Data

114

A Data Management Process Using SAS Enterprise Guide

Sporting Goods Case Study

Task: Add different types of source data to a process flow using SAS Enterprise Guide.

115

Explore the Data and Create a ReportInvestigate the products sold by country. Create a report of product category revenues by

country.

Which product categories sell the greatest quantity of products in each country? Which account for the highest revenue in each country? Create a graph of total quantity for each country. Create a graph of total revenue for each country.

116

Exploratory Analysis

117

Exploring the Data and Creating a Basic Report

Sporting Goods Case Study

Task: Create a tabular report of sales in each product category by country, a plot of number of products sold in each country and a plot of total revenue per country. Produce an HTML report to give to management.

118

IDEA EXCHANGE What conclusions would you draw from this basic data

exploration? Are there additional plots or reports you would like to explore from the orders data to help you determine where your marketing budget should be allocated?

What additional data would you need to help you make a case to the head of the Marketing Department that Marketing dollars should be spent in a particular way?

119

Chapter 1: Overview



1.3 Data Management


120

Recommended ReadingDavenport, Thomas H. and Jeanne G. Harris. “The Dark Side of Customer Analytics.” Harvard Business Review. May 2007. Plus commentary from George L. Jones, Katherine N. Lemon, David Norton, and Michael B. McCallister.

https://hbr.org/download/2707842/R0705A-PDF-ENG/R0705A-PDF-ENG.PDF

https://hbr.org/download/2707842/R0705A-PDF-ENG/R0705A-PDF-ENG.PDF

121

Recommended ReadingDavenport, Thomas H., Jeanne G. Harris, and Robert Morison. 2010. Analytics at Work: Smarter Decisions, Better Results. Boston: Harvard Business Press. Chapter 1, pp. 1-17; part 2 pp. 23-43 (overview and

data); pp. 19-22 optional

122

Recommended ReadingHaque, Umair. “The Case for Being Disruptively Good.” Harvard Business Review blogs. April 12, 2010. http://blogs.hbr.org/haque/2010/04/the_case_for_being_disruptivel.html

http://blogs.hbr.org/haque/2010/04/the_case_for_being_disruptivel.html

1 Chapter 1: Overview 1.1 Overview of Business Analytics 1.2 Software Used in This Course 1.3 Data Management 1.4 Recommended Reading.

Documents

analytics slide

useful analytics

data management

tons of data slide

good analytics

proliferation of data

overview of business

data collection protocol