1
DECISION SUPPORT TOOLS
Lecture 3: Multi-Dimensional Data Analysis
Module Outline
2
Management Decision-Making Multi-Dimensional Data Analysis
Stair, Reynolds & Chesney 124-125 & 219-221
Group and Executive Support Systems Model-based Decision Support Systems Intelligent Systems Knowledge Management Managing Decision Support Tools
Business Intelligence (BI)
What is Business Intelligence?
4
Business Intelligence is
A broad category of applications and technologies for gathering, storing, analysing, and providing access to data, to help enterprise users make better business decisions. BI applications include the activities of decision support systems, query and reporting, online analytical processing (OLAP), statistical analysis, forecasting, and data mining. Definition from TechTarget.com
Relate BI to decision support
5
Intelligence stage (of decision-making)
Becoming aware of the problem (or opportunity) Defining the problem details and scope
Design stage
Identifying alternative solutions Evaluating feasibility of each solution
Choice stage
Deciding on the best solution
Internal or external data? Actual or estimated values?
Knowledge Discovery
Multi-dimensional data analysis
8
OLAP and Data Mining Explores the relationship between multiple variables
Usually at least three variables involved Relies on large data sets Usually has a time component
Graphical display aids understanding
Identifies patterns occurring in data
Can provide a basis for developing mathematical models
On-Line Analytical Processing (OLAP)
9
Initiated by users Reveals relationships between data items Detects trends Clarifies problem definition Easy to use
Visual interface Flexible Drill-down
NOT done on operational database
Example of OLAP output
10
Slice and dice views
11
Different people may have different interests in the same dataset
OLAP vs standard DBMS queries
12
Quick to compose, run and modify No programming skills needed Visual output is more user friendly Key measures are already calculated OLAP structure can be used to build models (e.g. financial) Can provide input to other applications (e.g. performance management)
Equivalent Excel functions
13
Pivot tables Pivot charts Statistical analysis
Correlation Multiple regression Analysis of variance Cluster analysis
14
18 months & R10 million
15
The business value of Supply Chain Intelligence enabled by a Global Procurement Intelligence solution would be derived through:
The ability to identify, measure, manage and report procurement spend, price and consumption variances, trends and cost pressures across the group as well as by providing contract and vendor spend visibility; The ability to drive group cost optimisation through best practice sourcing strategies; Risk identification and mitigation capability; Visibility and standardisation of consolidated spend and related policies; and Data consolidation across the group, aligned to budgets and expenditure forecasts.
Moving on to Data Mining
16
Automated process to extract information from large data sets
Only as effective as the data it uses
Relies on advanced logical, statistical and mathematical techniques Infers rules and relationships that allow the prediction of future results
But cant explain underlying reasons
How is data mining being used in SA?
17
CRM and direct marketing Customer and product planning Fraud detection Credit scoring Benefits include:
Customer attraction and retention Product design Bad debt reduction Bank robbery prediction Product cross-selling
Comparing OLAP and data mining
18
Characteristic
Purpose
OLAP
Supports data analysis and decision making
Data Mining
Supports data analysis and decision making
Type of analysis Top-down, query-driven data Bottom-up, discovery-driven supported analysis data analysis Skills required of user Must be very knowledgeable Must trust in data mining about the data and its tools to uncover valid and business context worthwhile hypotheses
Web mining
19
Web Content Mining
Extraction of information from documents and databases
Web Structure Mining
Link structures within the Internet, most frequently visited paths etc
Web Usage Mining
Analyses data from the actions of Internet users (web and proxy server logs, user sessions, user profiles, registration data, cookies, bookmarks, mouse clicks etc.)
20
And thats all for today
