Week 5: Chapter 6 - Data: Business Intelligence

value of data

Trait	Definition/Purpose
Data Type	Data is categorized primarily as Transactional data (supporting daily operational/structured decisions) or Analytical data (supporting managerial analysis/semistructured decisions).
Data Timeliness	This refers to how current the data is, varying by industry. Real-time data is immediate and up-to-date data, provided by a real-time system in response to requests.
Data Quality	Business decisions are only as good as the quality of the data used to make them. Low quality results in data inconsistency (the same data element has different values) or data integrity issues (incorrect, inconsistent, or duplicate data).
Data Governance	This is the overall management of the availability, usability, integrity, and security of company data. It involves policies and procedures establishing rules for how data is organized, updated, maintained, and accessed. Master Data Management (MDM) is commonly included in data governance.

Data – Granularity, Format, Level

When making decisions, employees must be able to correlate data from various levels, formats, and granularities.

• Granularity: Refers to the extent of detail within the data.

  • Data can range from fine (detailed) to coarse (abstract/summary/aggregate).

  • The greater the granularity, the deeper the level of detail or fineness of data. Data can be captured in varying time intervals (e.g., seconds, minutes, hours).

• Levels: Data exists at the Individual, Departmental, and Enterprise levels.

• Formats: Data can exist in various formats, including Document, Presentation, Spreadsheet, and Database.

Characteristics of High Quality Data

High-quality data possesses five common characteristics: accuracy, completeness, consistency, timeliness, and uniqueness.

1. Accurate: The values in the data are correct (e.g., names are spelled correctly, dollar amounts are recorded properly).

2. Complete: No required values are missing from the data (e.g., an address includes street, city, state, and zip code).

3. Consistent: Aggregate or summary data agrees with the detailed data.

4. Timely: The data is current with respect to business needs (e.g., updated hourly, daily, or weekly).

5. Unique: Each transaction and event is represented only once in the data (e.g., no duplicate customers exist).

Relational DBMS

A database management system (DBMS) is software that manages and organizes large amounts of data efficiently.

• Definition: The DBMS creates, reads, updates, and deletes (CRUD) data in a database while controlling access and security. It provides a secure and reliable way to store data and ensures data integrity and consistency.

• Relational Model: The relational database model stores data in the form of logically related two-dimensional tables.

• Data Retrieval Tools: Managers use Query-by-Example (QBE) tools to graphically design answers to questions against a database. MIS professionals often use Structured Query Language (SQL), which requires writing lines of code to communicate with databases.

Data Elements and Attributes

Databases organize data using entities, attributes, and records.

• Entity (Table): Stores data about a person, place, thing, transaction, or event (e.g., CUSTOMER, PRODUCT). In the relational data model, entities become tables.

• Data Element (Field): The smallest or basic unit of data (e.g., a customer's name, product name). In the relational data model, attributes become fields.

• Attributes (Columns/Fields): The data elements associated with an entity.

• Record: A collection of related data elements, occupying one row in its respective table.

• Null-Valued Attribute: An attribute created when no other value applies or when a value is unknown (e.g., for a person without a mobile phone number).

Primary Key, Secondary Key (Foreign Key)

Keys are used to manage and organize various entities and create logical relationships.

• Primary Key: A field (or group of fields) that uniquely identifies a given record in a table. It is critical for distinguishing each record (e.g., a unique ID for Steve Smith, rather than just his name). A primary key must not contain the value null.

• Foreign Key: A primary key of one table that appears as an attribute in another table and acts to provide a logical relationship between the two tables. Matching foreign keys to primary keys represents the data relationships in a relational database.

Database Business Advantages

Relational databases offer several business advantages over simple spreadsheets or text documents:

• Increased Flexibility: Databases can support multiple ways for each user to access data, providing flexibility through the distinction between physical and logical views.

• Increased Scalability and Performance: Databases can handle massive volumes of data and large numbers of users quickly. Data latency (the time it takes for data retrieval/storage) must be minimized to maintain performance.

• Reduced Data Redundancy: Databases are designed to eliminate data redundancy (duplication of data) by recording each piece of data in only one place, which saves space and improves data quality.

• Increased Data Integrity (Quality): Integrity constraints (rules that ensure data quality), including business rules (defining how a company performs certain aspects of business), are enforced by the database and DBMS.

• Increased Data Security: Databases offer security features like passwords (authentication), access levels, and access controls to protect data from unauthorized users or misuse.

Database - Physical View

The distinction between physical and logical views is important for understanding flexible database access.

• The Physical view of data deals solely with the physical storage of data on a storage device.

• A database has only one physical view.

Database – Logical View and Section

Logical View: The Logical view of data focuses on how individual users logically access data to meet their particular business needs. For instance, one user might want an alphabetical report (Last Name, First Name) while another needs a mailing list (First Name, Last Name)—both are different logical views of the same physical data. A database can easily support multiple logical views.

Business Intelligence (BI)

helps managers solve the challenge of being "Data Rich, Information Poor". Many organizations struggle to understand their own strengths and weaknesses because the enormous volume of organizational data is inaccessible to most employees outside the MIS department.

• Organizational Data: This includes structured data from databases, as well as unstructured data such as voice mail, customer phone calls, text messages, video clips, and tweets.

• Data Aggregation (The Solution): BI uses data aggregation, which is the collection of data from various sources for processing, to compile data from multiple systems to support strategic analysis.

• BI Dashboards: Business intelligence dashboards track corporate metrics such as Critical Success Factors (CSFs) and Key Performance Indicators (KPIs). They include interactive controls that allow users to manipulate data for analysis. These dashboards offer analytical capabilities such as:

  • Consolidation: Aggregation of data from simple roll-ups to complex groupings (e.g., rolling up sales data from individual representatives to regional sales levels).

  • Drill-Down: Enables users to view details, and details of details, of information (the reverse of consolidation, allowing viewing of monthly data down to hourly data).

  • Slice-and-Dice: The ability to look at information from different perspectives (often along a time axis to analyze trends).

  • Pivot: Rotates data to display alternative presentations of the data.

Data Mining Analysis Techniques

Data mining is used to perform advanced analytics and discover patterns and relationships in data. Four common techniques are:

1. Estimation Analysis: Determines values for an unknown continuous variable behavior or estimated future value (e.g., predicting the percentage of students who will graduate based on income levels).

2. Affinity Grouping Analysis: Reveals the relationship between variables along with the nature and frequency of those relationships. This often uses Market Basket Analysis to detect customer buying behavior and identify affinities among choices (e.g., cross-selling related products).

3. Cluster Analysis: A technique used to divide an information set into mutually exclusive groups (clusters) where members are as close together as possible, and the groups are as far apart as possible (e.g., grouping similar customer attributes for targeted promotions).

4. Classification Analysis: The process of organizing data into predefined categories or groups (classes) for effective use (e.g., classifying customers who defaulted on a loan versus those who did not).