Data Science Case Study Prep

Case studies test your ability to apply data science to real business problems. Master structured thinking, problem decomposition, and clear communication.

Case Study Framework

Common Case Study Types

Practice Example: Churn Prediction

Problem: A SaaS company has 20% monthly churn. CEO wants to reduce it to 12%.

Step 1: Clarify

• What is the current customer base? (100K active)
• What defines churn? (No login for 30 days)
• What data is available? (usage logs, support tickets, billing)
• What is the budget for interventions? ($500K/yr)

Step 2: Structure

Root cause analysis
  ├── Why do customers churn?
  │   ├── Low engagement
  │   ├── Poor onboarding
  │   ├── Feature gaps
  │   └── Price sensitivity
  ├── Predict who will churn?
  │   ├── Feature engineering
  │   ├── Model selection
  │   └── Validation strategy
  └── How to intervene?
      ├── Retention offers
      ├── Product improvements
      └── Success outreach

Step 3: Analyze

# Feature engineering for churn
features = {
    "login_frequency_7d": "count(logins, last 7 days)",
    "session_duration_avg": "mean(session_length)",
    "support_tickets_30d": "count(tickets, last 30 days)",
    "days_since_last_login": "current_date - max(login_date)",
    "feature_adoption_rate": "features_used / total_features",
    "billing_age_days": "current_date - signup_date"
}

Step 4: Recommend

1. Short-term: Target top 10K at-risk customers with retention offers
2. Medium-term: Improve onboarding for first-30-day users
3. Long-term: Build real-time churn prediction system

Fermi Estimation Framework

Question: How many Google searches happen per day in India?

India population: 1.4B
Internet users: 50% = 700M
Daily search users: 60% = 420M
Searches per user: ~5/day
Total: 420M x 5 = 2.1B searches/day

Key Communication Tips

• Lead with the answer — state your recommendation first
• Use frameworks — structure shows systematic thinking
• Acknowledge trade-offs — every solution has costs
• Quantify everything — use numbers, not adjectives
• Ask clarifying questions — shows maturity

Key Takeaways

• Clarify before solving — ask smart questions
• Structure your thinking — use frameworks and trees
• Quantify impact — revenue, users, time saved
• Communicate clearly — executives need decisions, not details