Unit 1: Introduction to Business Intelligence — IT 249 Business Intelligence Study Notes

Pass path: Learn the definition of Business Intelligence (BI), the 4 stages of the analytical maturity curve (Descriptive, Diagnostic, Predictive, Prescriptive), and the high-level BI architecture. Top path: Contrast tactical operational reporting with strategic BI decision systems, evaluate real-time stream analytics vs batch processing, and analyze ROI calculation for enterprise BI investments.

Part 1: Core Concepts (2 Marks)

★ Business Intelligence (BI): An umbrella term that combines architectures, data integration tools, databases, analytical tools, and visualization applications to enable business users to transform raw operational data into actionable strategic insights.
★ The Four Tiers of Analytics:
1. Descriptive Analytics: What happened? (Historical reporting, KPI dashboards).
2. Diagnostic Analytics: Why did it happen? (Drill-down, root-cause OLAP).
3. Predictive Analytics: What will happen? (Statistical forecasting, machine learning models).
4. Prescriptive Analytics: What should we do about it? (Optimization algorithms, simulation models).
★ Key Performance Indicator (KPI): A quantifiable metric that reflects the critical success factors of an organization and measures progress toward specific operational or strategic goals.

Part 2: Structural Frameworks (10 Marks)

1. Business Intelligence Architecture: Framework, Value Chain, and Executive Decision Support

Introduction: Define BI. Explain that modern enterprises suffer from being “data rich but information poor” (DRIP syndrome). BI bridges this chasm by delivering the right information to the right decision-makers at the right time.

Comprehensive BI Architectural Pipeline

Data Sourcing Layer: Operational OLTP databases (ERP, CRM, e-commerce transactions), flat files, legacy systems, external web APIs, and IoT telemetries.
Data Integration / ETL Layer: Extracting raw data from disparate heterogeneous sources, Transforming data (cleaning, standardizing dates, deduplicating keys), and Loading into centralized stores.
Data Storage & Warehouse Layer: Enterprise Data Warehouses (EDW), Operational Data Stores (ODS), and departmental Data Marts modeled via dimensional schemas.
Analytics & Processing Layer: OLAP multidimensional analytical servers, data mining engines, and predictive statistical algorithms.
Presentation & Visualization Layer: Executive dashboards, automated threshold alerts, mobile BI apps, ad-hoc SQL query interfaces (Power BI, Tableau).

Importance of BI in Contemporary Decision-Making:
- Eliminates gut-feeling guesswork by anchoring decisions in empirical data.
- Compresses cycle times for operational reporting from weeks to seconds.
- Enables cross-departmental alignment around a Single Source of Truth (SSOT).
- Identifies emerging customer churn patterns before revenues collapse.
Conclusion: BI is not merely a software suite; it is an organizational capability converting historical data assets into enduring competitive advantage.

Part 3: Distinction Tables

Criteria	Operational Systems (OLTP)	Business Intelligence Systems (BI / OLAP)
Primary Objective	Run day-to-day core transactional operations smoothly	Support long-term management decision analysis
User Base	Frontline operational clerks, customers, cashiers	Executives, business analysts, data scientists
Data Nature	Current, real-time snapshot; highly volatile and updated in-place	Historical, summarized, non-volatile, periodic batch loads
Query Type	Simple, highly predictable, rapid read/write lookups	Complex, multi-table aggregations scanning millions of records
Database Design	Normalized (3NF) to maximize write speed and eliminate anomalies	De-normalized Star/Snowflake schemas optimized for analytical reads

Part 4: Worked Problems & Exam Traps

Problem 1 (Easy): Classify each of the following business questions into one of the four tiers of analytics (Descriptive, Diagnostic, Predictive, Prescriptive):
1. “Which 10 products generated the highest gross revenue last month?”
2. “Why did delivery times surge by 40% in our Biratnagar distribution hub in November?”
3. “What is the expected customer churn rate next quarter if competitor pricing drops by 5%?”
4. “What is the optimal truck dispatch route schedule to minimize fuel cost while meeting 100% of delivery deadlines?”
- Solution:
  1. Highest gross revenue: Descriptive Analytics (summarizes past events).
  2. Why delivery times surged: Diagnostic Analytics (analyzes root causes).
  3. Expected churn rate: Predictive Analytics (forecasts future probabilities).
  4. Optimal truck routing: Prescriptive Analytics (recommends optimal action).
Problem 2 (Exam-Level ROI Justification): How can a Chief Information Officer (CIO) quantitatively calculate the Return on Investment (ROI) for a Rs. 5,000,000 enterprise BI software implementation?
- Solution:
  1. Quantifiable Cost Savings (Direct Tangible Benefits):
    - Reduction in manual reporting labor: 5 data analysts saving 15 hours/week at Rs. 500/hr $= 5 \times 15 \times 52 \times 500 = \text{Rs. } 1,950,000$ annually.
    - Inventory holding cost reduction: Optimizing supply chains cuts dead stock holding costs by Rs. 2,000,000 annually.
    - Churn prevention: Early intervention saves Rs. 1,500,000 in customer lifetime revenue annually.
  2. Total Annual Tangible Benefits $= 1,950,000 + 2,000,000 + 1,500,000 = \text{Rs. } 5,450,000$ per year.
  3. Payback Period:
    $\text{Payback Period} = \frac{\text{Initial Investment}}{\text{Annual Benefits}} = \frac{5,000,000}{5,450,000} \approx 0.92 \text{ years (approx. 11 months)}$
  4. ROI over 3 years (assuming Rs. 500,000 annual maintenance):
    $\text{Net Benefit} = (5,450,000 \times 3) - (5,000,000 + 1,500,000) = 16,350,000 - 6,500,000 = \text{Rs. } 9,850,000$
    
    $\text{ROI} = \frac{9,850,000}{6,500,000} \times 100\% \approx 151.5\%$
Problem 3 (Twist / Case Scenario): A company installs Power BI licenses for all its 100 managers, but 6 months later, nobody uses the system and managers continue emailing around manual Excel spreadsheets. What went wrong?
- Solution:
  1. The organization treated BI as a pure software purchase rather than a cultural transformation.
  2. Common failure modes: Lack of data literacy training, missing executive sponsorship, dirty unvalidated underlying data that caused managers to lose trust in dashboard figures, and failure to decommission legacy Excel workflows.
  3. Successful BI adoption requires data governance, data stewardship, and change management.
- Students lose marks here because they define BI as just “installing Tableau”, or they confuse transactional OLTP databases with analytical BI systems.

Business Intelligence

Chapter notes

Unit 1: Introduction to Business Intelligence· 5 LHs· ~6 marks