Approaches to Big Data Analytics

1. Definition

Big Data Analytics is the process of examining large and varied datasets to uncover hidden patterns, correlations, market trends, customer preferences, and other useful information that can help organizations make informed business decisions.

---

2. Types of Big Data Analytics

Big Data analytics can be classified into four main approaches based on the nature and purpose of analysis:

Loading diagram…

---

3. Descriptive Analytics

Definition: Descriptive analytics examines historical data to understand what has happened in the past by summarizing and presenting data in meaningful ways.

Purpose: To provide insights into past performance and trends.

Techniques Used:

1. Data Aggregation: Combining data from multiple sources
2. Data Mining: Discovering patterns in large datasets
3. Statistical Analysis: Calculating means, medians, standard deviations
4. Visualization: Creating charts, graphs, and dashboards

Example Applications:

Industry	Application	Insight
Retail	Sales reports	Total sales by product category last quarter
Banking	Transaction history	Average daily transactions per branch
Healthcare	Patient records	Number of patients treated for specific diseases
E-commerce	Website analytics	Page views and bounce rates

Business Value:

• Provides baseline understanding of business operations
• Helps identify trends and patterns
• Supports data-driven decision making
• Enables performance benchmarking

---

4. Diagnostic Analytics

Definition: Diagnostic analytics examines data to understand the root causes of events and behaviors, answering "why" something happened.

Purpose: To identify factors that contributed to a particular outcome.

Techniques Used:

1. Drill-down Analysis: Examining data at granular levels
2. Data Discovery: Finding relationships and correlations
3. Correlation Analysis: Identifying relationships between variables
4. Root Cause Analysis: Determining underlying causes of problems

Example Applications:

Retail: Why did sales decrease in Q2?

• Analysis reveals: Product quality complaints increased
• Diagnostic finding: Supplier changed, quality declined
• Action: Switch back to original supplier

Banking: Why did loan defaults increase?

• Analysis shows: Defaults concentrated in specific geography
• Diagnostic finding: Local economic downturn
• Action: Tighten lending criteria for that region

E-commerce: Why did cart abandonment increase?

• Analysis indicates: Surge during checkout page redesign
• Diagnostic finding: New checkout process confusing
• Action: Redesign checkout for better usability

Business Value:

• Helps understand cause-effect relationships
• Enables targeted problem resolution
• Prevents recurrence of issues
• Improves operational efficiency

---

5. Predictive Analytics

Definition: Predictive analytics uses statistical models and machine learning algorithms to forecast future outcomes based on historical data.

Purpose: To anticipate future trends, behaviors, and events.

Techniques Used:

1. Regression Analysis: Predicting numerical values
2. Classification: Categorizing into predefined classes
3. Time Series Analysis: Forecasting based on time-ordered data
4. Machine Learning: Training models to make predictions
5. Neural Networks: Complex pattern recognition

Example Applications:

Retail - Demand Forecasting:

Historical sales data + Seasonal factors + Economic indicators
→ Predict: Next quarter's demand for each product
→ Action: Optimize inventory levels

Banking - Credit Scoring:

Customer data + Payment history + Employment status
→ Predict: Probability of loan default
→ Action: Approve/reject loan applications

Healthcare - Disease Prediction:

Patient records + Genetic data + Lifestyle factors
→ Predict: Risk of developing diabetes
→ Action: Preventive care recommendations

E-commerce - Customer Churn:

Purchase frequency + Website engagement + Customer service interactions
→ Predict: Likelihood of customer leaving
→ Action: Retention campaigns for at-risk customers

Business Value:

• Enables proactive decision making
• Reduces risks and uncertainties
• Improves resource allocation
• Enhances competitive advantage

---

6. Prescriptive Analytics

Definition: Prescriptive analytics recommends specific actions to achieve desired outcomes by analyzing possible scenarios and their consequences.

Purpose: To provide actionable recommendations and optimize decisions.

Techniques Used:

1. Optimization Algorithms: Finding best solutions
2. Simulation: Testing different scenarios
3. Decision Trees: Mapping decision paths
4. Linear Programming: Optimizing resource allocation
5. What-if Analysis: Evaluating impact of different choices

Example Applications:

Supply Chain:

• Problem: Minimize delivery costs while meeting deadlines
• Prescriptive Analysis:
  • Input: Fuel costs, traffic patterns, delivery windows, vehicle capacity
  • Output: Optimal routes for each delivery truck
  • Recommendation: Specific route assignments for today

Pricing Strategy:

• Problem: Maximize revenue for airline tickets
• Prescriptive Analysis:
  • Input: Demand forecasts, competitor prices, seat availability, time to departure
  • Output: Dynamic pricing recommendations
  • Recommendation: Set price at ₹5,500 for next 48 hours, then ₹6,200

Marketing Campaign:

• Problem: Allocate ₹1 crore marketing budget across channels
• Prescriptive Analysis:
  • Input: Historical ROI per channel, target audience preferences, seasonal factors
  • Output: Optimal budget distribution
  • Recommendation: Spend 40% on digital ads, 30% on TV, 20% on print, 10% on events

Business Value:

• Provides specific, actionable recommendations
• Optimizes complex decisions
• Automates decision-making processes
• Maximizes desired outcomes (revenue, efficiency, satisfaction)

---

7. Progression of Analytics Maturity

Loading comparison…

Organizations typically progress through these levels as their analytics capabilities mature.

---

8. Comparison of Analytics Approaches

Approach	Question	Time Focus	Techniques	Business Impact
Descriptive	What happened?	Past	Reports, dashboards	Understanding
Diagnostic	Why happened?	Past	Correlation, drill-down	Problem solving
Predictive	What will happen?	Future	ML, forecasting	Planning
Prescriptive	What to do?	Future	Optimization, simulation	Action

---

Exam Pattern Questions and Answers

Question 1: "Explain the four types of Big Data Analytics with suitable examples." (10 Marks)

Answer:

Introduction (1 mark):
Big Data Analytics can be classified into four types based on the purpose and nature of analysis: Descriptive, Diagnostic, Predictive, and Prescriptive Analytics.

Descriptive Analytics (2 marks):
Descriptive analytics examines historical data to understand what happened in the past by summarizing and presenting data in meaningful ways. It uses techniques like data aggregation, data mining, and statistical analysis. For example, a retail company analyzing total sales by product category for the last quarter to understand past performance.

Diagnostic Analytics (2 marks):
Diagnostic analytics examines data to understand root causes of events, answering why something happened. It employs techniques like drill-down analysis, correlation analysis, and root cause analysis. For example, an e-commerce company investigating why cart abandonment increased and discovering that a recent checkout page redesign made the process confusing.

Predictive Analytics (2.5 marks):
Predictive analytics uses statistical models and machine learning to forecast future outcomes based on historical data. It employs techniques like regression analysis, time series forecasting, and neural networks. For example, a bank using customer data and payment history to predict probability of loan default, enabling better credit decisions.

Prescriptive Analytics (2.5 marks):
Prescriptive analytics recommends specific actions to achieve desired outcomes by analyzing scenarios and consequences. It uses optimization algorithms, simulation, and decision trees. For example, an airline using dynamic pricing algorithms that recommend specific ticket prices based on demand forecasts, competitor prices, and seat availability to maximize revenue.

---

Question 2: "Differentiate between Predictive and Prescriptive Analytics." (4 Marks)

Answer:

Predictive Analytics (2 marks):
Predictive analytics forecasts what will happen in the future based on historical patterns. It provides predictions and probabilities of future events. For example, predicting next quarter's sales or customer churn probability. It answers "What will happen?" and helps in planning and preparation.

Prescriptive Analytics (2 marks):
Prescriptive analytics recommends what actions should be taken to achieve desired outcomes. It goes beyond predictions to provide specific recommendations. For example, recommending optimal price points or best delivery routes. It answers "What should we do?" and directly supports decision-making and optimization.

---

Summary

Key Points for Revision:

1. Descriptive Analytics: What happened? (historical analysis)
2. Diagnostic Analytics: Why did it happen? (root cause analysis)
3. Predictive Analytics: What will happen? (forecasting)
4. Prescriptive Analytics: What should we do? (recommendations)
5. Progression: Organizations move from descriptive → diagnostic → predictive → prescriptive
6. Value Increase: Each level adds more business value and complexity

---

Quiz Time! 🎯

Loading quiz…