Information Technology and Modern Business Organizations

What is the Total Quality Management Approach?

Total Quality Management Approach

a. Define –

b. Measure –

c. Analyze –

d. Improve -

 (Accuracy, Integrity, Consistency, Completeness, Validity, Timeliness, Accessibility)

Business Intelligence – use of technology and statistical techniques to gather, analyze large amounts of data to support business making decisions.

 Data Mining – a common technological / computational approach to extract business intelligence from vast amounts of data.

Data mining – process that extracts previously unknown interesting, valid and actionable data

patterns from a large set of data.

Data Mining Process >

a. Business understanding or ask question – define what the information will be used for and what the mining will be used for.

 b. Data understanding / Selection – identify the appropriate and most relevant data from various databases.

 c. Data Preparation – prepare data for data mining analysis. Relevant data is collected, then cleaned / scrubbed (remove outliers, inconsistencies), then is transformed / normalized (consistent values) and data reduction (remove data that will not be needed or is not applicable)

 d. Model Building – various modeling techniques are applied to the data to answer the question. This includes determining the best technique.

 e. Testing and Evaluation – models are assessed for accuracy analyzes what model meets the business objective.

 f. Interpretation / Deployment – determine how to present the information to the end user and interpret the results.

Data Mining Patterns

Associations – commonly co-occurring groupings of things such as beer or diapers

 Predictive – tell the nature of the future

 Clusters – Identify natural groupings of things based on characteristics suchs as assigning customers into different segments

 Sequential relationships – discover time-ordered events such as predicting an existing banking customer who already has checking and will open an investment account within a year

Support ( X – Y) = n (X and Y) / N (total number)

 Confidence (X – Y) =. n (X and Y) / n(X) number of transactions in data that contain both

X and Y and total number of items that include only X

 1. Data mining tool can discover groupings within the data such as finding affinity groups or partitioning a database into groups of customers based on demographics.

2. Divides a data set into mutually exclusive distinct groups such that members of each group are as close together to one another and different groups are as far apart as possible.

 3. Unlike classification the class labels are unknown

 4. Most common types are k-means and self-organizing

 5. E.g. works in manner similar to classification when no groups have been defined.

Classification

Most Common

Recognizes patterns that describe the group to which an item belongs by examining exiting items that have been classified by inferring a set of rules.

Most common type of data mining

 1. Analyze historical data to predict the future behavior

2. Assign an instance to predefined outcome classes.

Used for predictive purposes

 E.g - Recognizes patterns that describe the group to which an item belongs by examining existing items that have been classified and by inferring a set of rules.

Common techniques are neural networks and decision tree analysis.

1. Redundancy

2. Quality

3. Identify

-Traditional file processing had data redundancy and data inconsistency (differing

conventions)

remove redundancy, increase data sharing, enhance data quality, improve data accessibility, reduce maintenance costs, reduce unplanned data

New staff, installation complexity, 

To design a database you need to understand the relationships among the data, the type of data that will be maintained in the database, how data will be used, how organization will need to change to manage data.

The design requires considering a conceptual design and physical design.

Conceptual Design – how data elements are to be grouped

Physical Design – determine how the database will be actually rearranged on direct-

access storage devices.

a. We start with requirements analysis (include data requirements and functional requirements (what queries we want). Requrements are the input to the first design stage

b. First Stage – conceptual design stage, the output is the ER (ER diagram)

c. Logical Design – output is the relationship scheme (output ect.)

1) Data Requirements – Define Data Requirments

2) Conceptual Database design – how data elements are going to be grouped

3) Physical design – how database will be actually organized on direct access storage devices

4) Database implementation

Business Process Management

Be able to discuss why your company should model

and document every business process. Be able to discuss how your company can

measure, monitor and improve it

Data flow diagram is process of modeling four constructs to model business processes:

1. Data flow

2. Data store

3. Process

4. External entity

Process – produce specific result, designed for targeted customers, initiated in response to an event, consisted of set of inter-related activities.

Data Flow diagrams

Level 0

Data Flow diagrams are hierarchical in structure and can be decomposed to represent additional levels.

Data store – data within the business (rectangle)

-Process – square with rounded edges (work or activity performed)

-External entity – origin or destination of data (source data sink) rectangle

- Process cannot have only input data or only output data

- Data flow directionally.

External Entity

Understand functional decomposition and balancing

May give a scenario and ask to produce level 0

-Level 0 – represent the major focal business processes

IT Investment Management

Limits to only using financial metrics

Developing a multi-criteria base to develop an investment

*The use of financial methods to evaluate / prioritize competing IT projects / requests

*Limitations? Why?

IT Benefits –

-Direct vs. Indirect –

(It is important to show the improvement and try to measure the outcome)

IT -> Improvement -> "Measurable" Outcome

Qualitative strategy (Customer experience, improved job satisfaction)

vs.

Quantitative (reduced wages)

(A lot of IT benefits are Indirect and Qualitative)

-Nature of big data – volume, velocity and variety

-Information Pyramid – Data > Information > Knowledge

-Database > File > Record > Field > Byte > Bit

-Entity – person, place or thing which we maintain information

-Attribute – characteristic describing the data

-Database - Collection of data organized to serve many applications efficiently by centralizing data and controlling redundant data.

-Logical view – presents data how end users or business specialists would see

-Physical view – how data are actually structured or organized

-Relational Database – two dimensional tables

-Primary Key – unique identifier for information in any row of the table.

-Foreign Key – primary key from another table

-Instance of – relationship between entities

a. Financial perspective

b. Customer perspective

c. Innovation and learning

d. Internal Business Perspective

a. Categorical data – labels to divide into groups

b. Nominal data – contain measurements of simple codes not measurable (e.g. married vs single)

c. Numeric data – represent numerals of specific data

d. Interval data – scale (e.g. temperature)

e. Ratio data – include measurement variables

Item set, Cardinality, Support of Item Set

Defined

Item set – set of items (e.g. eggs, milk, etc.)

-Cardinality – exact number of items in an item set

-Support of item set – ratio between number of transactions that include all items in

item set and total number of transactions under analysis.

Cardinality constraint

specifies exact number of instances participating in the entity set

- One to one

- One to many

- Many to one

- Many to many

Process Measurements

1. Time

2. Quality

3. Cost

4. Customer perspective