Managing Information Systems

Managing Information Systems Enhancing Management Decision Making Part 1 Section 13.1

1

Objectives • To understand types of decision-support systems • To understand the components of a decision-support system

2

Decision-support Systems • What is a decision-support system (DSS)?

3

MIS or DSS? • Management Information Systems: – Routine reports (periodic) – Assist control of an organisation

• Decision-support Systems: – Non-routine – Support flexibility and rapid response – Semi-structured or unstructured data 4

Types of DSS • Model-driven – – – –

Uses a model to perform ‘what if’ analysis Typically standalone In-house or departmental Strong theory or model

5

Types of DSS • Data-driven – Analyse large amounts of data – Data from TPS into data warehouses – Use • On-line Analytical Processing (OLAP) • Data mining

6

Data-driven Examples • Contrast – How many widgets were shipped in December?

• With – Compare the sales of widgets to the sales plan by quarter and sales region for the last two years?

7

DSS Components TPS

User Interface

DSS Database

External Data

DSS Software System: Models OLAP Tools Data Mining Tools

User 8

DSS Models • Abstract representation that illustrates the components or relationships of the problem – Physical: model of an airplane – Mathematical: profit = revenue - costs – Verbal: description of a procedure

9

DSS Models • • • •

Statistical (typical) Optimisation Forecasting Sensitivity analysis – “What if” – Repeatedly modify parameters of model to determine outcome 10

OLAP (On-line Analytical Processing) • Dynamic multi-dimensional analysis of enterprise data • Just-in-time information • Wide variety of views of information • Transformation of raw data: – Reflects the ‘real’ dimensionality of enterprise

11

OLAP • Data: – Loading – bulk and operational, internal and external – Aggregation

• Processing: – Application of business models and statistics

• Querying: – Complex – Drill-down through hierarchies – Ad-hoc 12

Data Mining • Provides a way of finding hidden insight not obtained by traditional techniques. • Uses: – – – – –

Statistical analysis Neural networks Fuzzy logic Genetic Algorithms Rule-based systems 13

Data Mining • Associations – Occurrences linked to a single event

• Example – Supermarket purchases – When crisps are bought, 85% of the time a can of Coca-cola is bought

14

Data Mining • Sequences – Events linked over time

• Example – House purchase – Within two weeks, 65% of the time a refrigerator is bought – Within one month, 45% of the time an oven is bought 15

Data Mining • Classification – Recognise pre-defined patterns to group similar items

• Example – Telephone operators – Recognise those attributes of customers who are likely to leave 16

Data Mining • Clustering – Recognise patterns to cluster similar items without pre-defined groups

• Example – Bank customer details – Partitioning data into groups by demographics or investments 17

Data Mining • Forecasting – Use existing data to forecast future values

• Example – Past performance to predict sales figures

18

DSS Examples • Supply Chain Management – Who, what, when and where? – Purchasing, manufacture and distribution

• Customer Relationship Management – Pricing – Customer retention – New revenue streams 19

DSS Examples • Business Scenarios – Sensitivity analysis of business parameters – Cost / benefit analysis

• Geographic Information Systems (GIS) – Display information geographically – Demographics, customers, crime

20

Example Questions 1. 2. 3. 4.

Who are our most frequent customers? Do they live close to our shops? How can we resegment those customers? How can we better reach those segments?

1. Customer data warehouse •Legacy data •Website transactions •Call centre data •External data

2. 3. 4.

Analysis Use statistical analysis to find top 25% most frequent customers. Establish correlation between location and sales Verify new customer segments Query database on customer information per segment 21