Data Science Interviews: Business Impact Through Analytics 2026

1. 1. Write a query to find the second highest salary from an employee table.

   Classic SQL question testing LIMIT, ORDER BY, and subqueries

2. 2. What's the difference between INNER JOIN and LEFT JOIN?

   Understanding data relationships and handling missing values

3. 3. How do you handle NULL values in SQL?

   COALESCE, ISNULL, NULL impact on aggregations

4. 4. Find duplicate records in a table.

   GROUP BY, HAVING, COUNT() for data quality checks

5. 5. Calculate running totals using SQL.

   Window functions and cumulative calculations

1. 6. Design a SQL query to calculate customer churn rate.

   Business metrics calculation with date functions

2. 7. How would you optimize a slow SQL query?

   Indexing, query execution plans, avoiding N+1 queries

3. 8. Write a query for cohort analysis.

   Date grouping, retention metrics, complex joins

4. 9. Find the top 3 products by revenue in each category.

   Window functions with RANK() and PARTITION BY

5. 10. Calculate month-over-month growth rate.

   LAG/LEAD functions for period comparisons

1. 11. Design a data warehouse schema for e-commerce analytics.

   Star schema, fact/dimension tables, data modeling

2. 12. How would you handle slowly changing dimensions?

   SCD Type 1, 2, 3 for historical data tracking

3. 13. Implement incremental data loading strategy.

   ETL optimization, change data capture, performance

4. 14. Design a real-time analytics dashboard query.

   Streaming data, materialized views, caching strategies

5. 15. How do you ensure data quality and consistency?

   Data validation, anomaly detection, governance frameworks

1. 16. What's the difference between mean, median, and mode?

   Central tendency measures and when to use each

2. 17. Explain the Central Limit Theorem in simple terms.

   Foundation for statistical inference and hypothesis testing

3. 18. What does p-value mean?

   Probability of observing results if null hypothesis is true

4. 19. How do you detect outliers in a dataset?

   IQR method, z-score, visual methods like box plots

5. 20. What's the difference between correlation and causation?

   Relationship types and avoiding analytical pitfalls

1. 21. How would you design an A/B test for a new feature?

   Sample size calculation, randomization, metric selection

2. 22. What's Type I vs Type II error?

   False positives vs false negatives, business implications

3. 23. Explain confidence intervals.

   Uncertainty quantification, interpretation pitfalls

4. 24. How do you handle multiple hypothesis testing?

   Bonferroni correction, False Discovery Rate

5. 25. What's the difference between Bayesian and frequentist statistics?

   Philosophical approaches to probability and inference

1. 26. How would you measure the impact of a marketing campaign?

   Causal inference, difference-in-differences, matching methods

2. 27. Design an experiment with network effects.

   Cluster randomization, spillover effects, social networks

3. 28. How do you handle seasonality in time series analysis?

   Decomposition, detrending, ARIMA models

4. 29. Explain bootstrapping and when you'd use it.

   Resampling methods for uncertainty estimation

5. 30. How would you model customer lifetime value?

   Cohort modeling, churn prediction, revenue forecasting

1. 31. What's the difference between supervised and unsupervised learning?

   Learning paradigms and problem types

2. 32. Explain bias-variance tradeoff.

   Model complexity, overfitting, underfitting

3. 33. How do you evaluate a classification model?

   Accuracy, precision, recall, F1-score, ROC curves

4. 34. What's cross-validation and why use it?

   Model evaluation, reducing overfitting, k-fold CV

5. 35. Explain linear regression assumptions.

   Linearity, independence, homoscedasticity, normality

1. 36. How would you handle imbalanced datasets?

   SMOTE, undersampling, cost-sensitive learning

2. 37. What's the difference between Random Forest and Gradient Boosting?

   Ensemble methods, bias-variance characteristics

3. 38. How do you select features for a model?

   Filter, wrapper, embedded methods, domain knowledge

4. 39. Explain regularization in machine learning.

   L1/L2 regularization, preventing overfitting

5. 40. How would you build a recommendation system?

   Collaborative filtering, content-based, hybrid approaches

1. 41. How do you deploy ML models in production?

   MLOps, model versioning, monitoring, A/B testing models

2. 42. How would you handle model drift?

   Data drift detection, model retraining strategies

3. 43. Explain model interpretability techniques.

   SHAP, LIME, feature importance for business stakeholders

4. 44. How do you scale ML training for large datasets?

   Distributed computing, mini-batch training, cloud ML platforms

5. 45. Design an ML system for fraud detection.

   Real-time inference, feature engineering, false positive management

1. 46. How would you investigate a 20% drop in user engagement?

   Root cause analysis, metric decomposition, hypothesis testing

2. 47. Design a metric to measure product success.

   North Star metrics, leading/lagging indicators, stakeholder alignment

3. 48. How would you prioritize features for development?

   Data-driven prioritization, impact estimation, resource constraints

4. 49. Estimate the business impact of a new recommendation algorithm.

   Revenue modeling, user behavior analysis, experimentation design

5. 50. How would you explain machine learning to a CEO?

   Business value communication, ROI justification, strategic alignment

The STAR-D Method for Data Science

Adapt the STAR method for technical interviews by adding "Data":

1. Situation: "The company was seeing declining user retention..."
2. Task: "I needed to identify the root cause and recommend solutions..."
3. Action: "I started by analyzing user behavior data, then segmented users by cohort..."
4. Result: "We discovered the issue was in onboarding, leading to a 15% improvement..."
5. Data: "I used SQL for data extraction, Python for analysis, and presented findings in Tableau..."

Common Interview Mistakes to Avoid