Business Intelligence and Data Science Overview

Study Notes

Business intelligence (BI) was previously everything related to using data for managerial decision support
Now, it's a part of business analytics
BI is descriptive analytics
Descriptive analytics answers "what happened?" and "what is happening?"
Predictive analytics answers "what will happen?" and "why will it happen?"
Prescriptive analytics answers "what should I do?" and "why should I do it?"
Business analytics uses BI, dashboards, scorecards, data warehousing and well-defined business problems/opportunities, to support Business Intelligence

A data warehouse is a physical repository for relational data, specially organized for enterprise-wide, cleansed data
Standardized formats
Central repository of information analyzed for informed decisions
Data flows into the data warehouse from transactional systems, relational databases, and other sources
Typically on a regular pace
The data warehouse is a collection of integrated, subject-oriented databases designed to support decision support systems (DSS) functions
Each unit of data is non-volatile and relevant to a specific moment in time

A relational database is a collection of information organized into predefined relationships
Data is stored in tables with columns and rows
Relationships between tables allow analysis of how different data structures relate to each other.

Subject-oriented: Data is organized by subject (e.g., sales, products, customers)
Integrated: Data from different sources is integrated, resolving naming conflicts and discrepancies
Time-variant: Contains historical data (e.g., daily, weekly, monthly, annually) for analysis of trends, deviations, long-term relationships, and forecasting
Non-volatile: Data cannot be changed; changes are recorded as new data, rather than updating the original data
Metadata: Data about data, describing how data is organized
Web based, relational/multi-dimensional client/server, real-time/right-time/active: Ensures easy access to end-users

A smaller, departmental data warehouse
Stores only limited/relevant data
Dependent data mart: Created directly from an existing data warehouse; all users view the same data version
Independent data mart: Lower cost, scaled-down version of a data warehouse for a strategic business unit or department; its source is not the enterprise warehouse

A type of database used as an interim/short-term storage area for a data warehouse
Provides a fairly recent form of customer information
Data is updated throughout the course of business operations
Used for short-term decisions
Classified as a short-term memory that stores very recent data; the data warehouse stores permanent data

Dimensions are perspectives/entities about which an organization keeps records (e.g., time, location, item type)
A 2D representation of data with two dimensions (organized in quarters and item) (e.g., sales of various items in a city per quarter)
A 3D representation where a third dimension (organization location) is added (e.g., items sold per quarter per city)

Three-tier architecture: Data acquisition software (back-end), data warehouse, client software (front-end)
Two-tier architecture: Combines the first two tiers in a three-tier architecture
One-tier architecture: Only one single tier physically accessible

Processes to integrate data from various business systems
Key tasks include data extraction, transformation, and loading
Uses Enterprise Application Integration (EAI)
Creates a process to deliver data to the data warehouse, to create a near-real-time data warehouse or to the OLTP systems

A major project that influences various organizational units/departments
Has direct benefits (enhanced analysis capabilities, consolidated data views, enhanced business performance, simplified data access); and indirect benefits (enhanced business knowledge, competitive advantage, improve customer service, facilitate decision-making)