ABW501 (Analytics Edge) Final Exam Review: Full Prediction Guide

📚 Exam Info

• Format: 4 structured questions
• Coverage: Weeks 1, 2, 5, 7 + core analytics concepts
• Tip: Keep answers concise; use tables/keywords; calculator allowed

---

📸 Visual Quick Reference (Henry Notes)

Images from Henry’s review pack.

Page 1

Page 2

Page 3

---

✅ Predicted 4 Major Question Types (Week 1/2/5/7)

Q1: Concepts of Analytics (Chapter 1)

Q1(a) Four Types of Analytics

Type	Key Idea	Question Answered	Example
Descriptive	Past facts	What happened?	Monthly sales report
Diagnostic	Root cause	Why did it happen?	Churn analysis
Predictive	Probabilities	What will happen?	Demand forecast
Prescriptive	Optimization	What should we do?	Optimal pricing

Answer tip: Use a table and keep it short.

Q1(b) BA vs DS Skill Focus

Skill	BA	DS	Key Difference
Business Knowledge	Very High	High	BA is business-heavy
Data Visualization	Very High	High	BA focuses on reporting
Statistics/Math	Some	Very High	DS is math-heavy
Machine Learning	None/Low	Very High	DS focuses on modeling
Data Wrangling	Some	Very High	DS spends more time on prep

Summary: BA = business communication; DS = modeling and statistics.

Visual: Analytics Types vs Roles

---

Q2: Analytics Lifecycle (Chapter 2)

Q2(a) Six Phases + Key Tasks

1. Discovery – define business problem, assess resources
2. Data Preparation – ETL, clean data, set up sandbox (most time-consuming)
3. Model Planning – select techniques & features
4. Model Building – train/test models
5. Communicate Results – visualize and interpret
6. Operationalize – deploy and monitor

Tip: Emphasize Phase 1 & 4 if asked.

Q2(b) Seven Key Roles (bonus)

1. Business User – provides problem context
2. Project Sponsor – budget & alignment
3. Project Manager – timeline & coordination
4. BI Analyst – reporting/visualization
5. DBA – data governance
6. Data Engineer – pipelines & prep
7. Data Scientist – modeling & insights

Visual: Lifecycle & Roles

---

Q3: Regression Analysis (Chapter 5)

Regression Types by Shape

1. Simple Linear Regression – straight line, continuous outcome
2. Logistic Regression – S-curve, binary probability
3. Polynomial/Non-linear – curve for non-linear relationships

Q3(a) Interpret a Regression Equation

Given $Sales = 100 + 5(Advertising)$:

• Intercept (100): predicted sales when advertising is 0
• Slope (5): each 1 unit increase in advertising raises sales by 5

Q3(b) Meaning of $R^2$

If $R^2 = 0.85$:

• the model explains 85% of variance in the dependent variable
• higher $R^2$ = better fit (but not necessarily causation)

---

Q4: Data Mining Algorithms (Chapter 7)

Common Algorithm Comparison

Algorithm	Best Use	Strengths	Weaknesses
Decision Tree	Interpretability	Easy to explain	Overfitting risk
Random Forest	Accuracy	Robust, lower variance	Less interpretable
KNN	Small data	Simple	Slow, sensitive to noise
Naive Bayes	Text/classification	Fast, low data requirement	Independence assumption
SVM	High-dimensional	Strong margins	Computationally heavy

Key rule: Match algorithm to data size, dimensionality, and explainability needs.

---

Final Tips

• Use tables and keywords in answers
• Keep definitions short, focus on comparison
• Use diagrams when possible
• Calculator is allowed for simple ratios/percentages