DDR Data Science Interview Practice Test

These questions are part of the practice quiz. Start practicing

• How would you determine the required sample size for an A/B test with a binary outcome, given baseline conversion rate, detectable uplift, significance level, and power?
• What is a hash table?
• Describe the Difference-in-Differences (DiD) approach and its identifying assumption.
• What is a Self-Join?
• In PostgreSQL, what does date_trunc('quarter', current_date) return?
• What is the difference between a list and a tuple in Python?
• What is the purpose of preprocessing steps like scaling, normalization, and one-hot encoding in ML pipelines?
• How would you communicate model findings to a non-technical stakeholder, focusing on business impact and uncertainties?
• What are the wildcard characters used with the LIKE operator?
• What are aggregate functions?
• What is a subquery?
• Which is an example of embedding in MongoDB for a one-to-many relationship?
• What is the difference between a shallow copy and a deep copy in Python?
• Which statement best describes the difference between UNION and UNION ALL?
• What is the purpose of SQL constraints?
• What is the purpose of cross-validation, and how would you implement 5-fold CV for regression?
• Name at least three common pitfalls in data visualizations and how to avoid them.
• How does a UDF differ from a stored procedure?
• Explain the concept of regularization strength and its effect on bias-variance tradeoff.
• What does TCL manage?
• In database normalization, what does Third Normal Form require?
• Which scenario would most likely benefit from using a stored procedure?
• What constraint enforces a relationship between two tables?
• In a confusion matrix, how do you compute sensitivity and specificity, and what do they indicate?
• What does a plateau in a learning curve suggest about data sufficiency and model capacity?
• Which statement correctly contrasts embedding and referencing in MongoDB?
• Which subquery type provides a temporary table-like result to be used in the FROM clause?
• How does MongoDB differ from a relational database?
• Which statement BEST describes the bias-variance trade-off in model selection?
• What are the limitations of views?
• What is the difference between BETWEEN and IN in SQL?
• Why are hash tables efficient for lookups?
• What is the difference between a parametric model and a flexible non-parametric model in terms of assumptions and capacity?
• Which SQL sublanguage defines schema?
• Why are cursors generally discouraged?
• Which approach can fetch alternate rows from a table?
• In MongoDB, embedding is storing related data as a nested document. Which option describes embedding?
• Define the concept of overfitting and how to detect it in practice.
• Relational database design principle?
• If a p-value is 0.04 and alpha is 0.05, what decision should you make?
• Describe how you would detect data leakage and provide an example in a churn prediction pipeline.
• Which of the following are examples of aggregate functions?
• What is the difference between forecasting and predicting in statistical terms?
• In pandas, which approach fills numeric columns with their median and categorical columns with their mode in a single pass?
• Which statement correctly describes batch, stochastic, and mini-batch gradient descent and when you would use mini-batch?
• What are ARIMA(p,d,q) components and how would you choose p, d, q?
• Describe regularization paths for L1 vs L2 penalties and their impact on feature selection.
• How do DELETE and TRUNCATE differ in behavior?
• Given a linear regression with multicollinearity (VIF > 5) among predictors, what issues arise and how might you address them?
• What is a database?
• Which statement about UDFs' ability to modify data is true?
• What is the difference between parametric and non-parametric models? Provide examples.
• What is the difference between WHERE and HAVING in SQL?
• Which SQL clause can fetch only distinct records from a table?
• Explain the Poisson distribution and a typical scenario where it models counts.
• Which statement best captures the modeling of one-to-many relationships in SQL versus MongoDB?
• Which of the following is an example of a non-parametric method in survival analysis?
• Which statement best describes the Bonferroni correction versus the Benjamini-Hochberg procedure, and when would you use each?
• Which scalar function returns the current date and time?
• Which is an advantage of stored procedures?
• Which statement is true about NoSQL data models?
• How would you model a one-to-many relationship in SQL?
• What best describes the effect of high variance on model performance as more data becomes available?
• What is the AUC-ROC curve and what does it measure? Why might PR-AUC be preferred in imbalanced problems?
• How would you approach feature engineering for a customer churn prediction problem?
• Which operator is used for pattern matching?
• MongoDB stores data in which format?
• Which function extracts the first N characters from a string in SQL Server or MySQL?
• What are the four basic types of SQL joins and their behavior?
• Which method is commonly used to estimate how well a model generalizes to unseen data?
• Which constraint ensures a column cannot be NULL?
• In survival analysis, what is the hazard function and what does the Cox proportional hazards model assume?
• Which of the following is NOT a type of subquery?
• Which of the following is a requirement of First Normal Form (1NF)?
• Batch processing mode processes data how?
• How should MNAR missingness be handled?
• Which NoSQL approach is commonly used to scale across many machines?
• What is feature importance in tree-based models, and how can SHAP values improve interpretability?
• How would you handle missing data if the missingness is completely at random (MCAR) vs. not at random (MNAR)?
• One-hot encoding is used in ML pipelines primarily to handle which type of features?
• Which statement best describes a scalar subquery?
• Which SQL construct would correctly include categories with zero revenue when calculating total revenue by product category for the last calendar quarter in PostgreSQL?
• What is the difference between ETL and ELT, and in which scenarios would you prefer each?
• Which pandas expression selects numeric columns from a DataFrame?
• Difference between SQL and MySQL?
• In which location is a subquery NOT listed as a valid location?
• Which statement about a unique key is true?
• What is a foreign key?
• Which statement about relationships between tables is true?
• What is a unique key?
• Which expression computes a 7-day running total of daily revenue in SQL using window functions?
• What happens when a view is accessed?
• Which of the following is an example of a scalar function?
• What is the purpose of regularization in logistic regression, and how does it affect coefficients?
• What does the USING clause do in a JOIN?
• What is a p-value and what does a small p-value in hypothesis testing indicate?
• In MongoDB terminology, which element maps to a SQL database?
• What is the Central Limit Theorem and why is it important in data science practice?
• Explain the difference between RMSE and MAE as error metrics.
• In Oracle, which function extracts the first N characters from a string?
• In SQL window functions, which frame specification yields a 7-day running total when ordering by date?
• In time-series data, what is stationarity, and which tests would you perform to assess it?
• What can happen to an index over time if it is not maintained?
• Which function converts a string to upper case?
• What is a View in SQL?
• Which condition defines Second Normal Form (2NF)?
• What is a lift chart and how is it used in evaluating model performance for marketing campaigns?
• Briefly explain TF-IDF and why it is useful in text classification.
• Which type of dependency is eliminated by Third Normal Form?
• Which statement best describes the relationship between a 95% confidence interval and a 5% significance level in hypothesis testing?
• In the Cox proportional hazards model, what does the hazard ratio represent?
• In MongoDB terminology, which concept maps to a SQL table?
• Define concept drift and describe a practical strategy to detect and respond to it in production models.
• Which join returns all rows from the left table and matching rows from the right table?
• In SQL, which join returns all rows from both tables, with NULLs where there is no match?
• Explain the bias-variance decomposition for mean squared error in a regression problem.
• Which operation triggers maintenance of an index?
• In MongoDB, embedding stores related data as a nested document. Which of the following options best describes embedding?
• Which of the following is true about using GROUP BY to fetch distinct records?
• In MongoDB, which statement best contrasts embedding and referencing?
• Which of the following are common trade-offs of using indexes?
• Which three components comprise RFM features commonly used in customer analytics?
• Which is an example of referencing in MongoDB for a one-to-many relationship?
• Which statement is true about stored procedures and security?
• What is a primary advantage of MongoDB over a relational database?
• Define Type I and Type II errors with examples in A/B testing.
• Which diagnostic plots are commonly used to choose ARIMA p and q values?
• In MongoDB, one-to-many modeling, which option corresponds to embedding the child documents in the parent?
• For MAR missingness, which imputation approach is commonly recommended?
• What is an RDBMS?
• What is SQL?
• Which metric is more informative for imbalanced datasets: PR-AUC or ROC-AUC?
• Which of the following is a primary reason to use mini-batch gradient descent instead of full batch gradient descent?
• In k-fold cross-validation for classification, how many models are trained and evaluated, and why is stratification important?
• Which statement about aggregate functions and GROUP BY is true?
• What is a Cross-Join?
• What is the primary reason indexes speed up queries?
• Main sublanguages of SQL?
• What is an effective approach to avoid overplotting in a dense scatter plot?
• Which join type returns all rows from both tables, including unmatched rows?
• What does collation define?
• What is a Python lambda function?
• To model a one-to-many relationship in SQL, which option uses a foreign key and a join to retrieve data?
• What is a cursor?
• Table, record, field definitions?
• Which scenario best favors choosing a NoSQL database over a relational database?
• In an imbalanced dataset with 2% positives, which evaluation metric is most informative for assessing performance on the minority class?
• What is the time complexity for quicksort, mergesort, and a hash-based lookup in Big-O terms?
• What is normalization in relational databases?
• Which of the following is a characteristic of non-parametric models?
• What is a primary key?
• What is a DBMS?
• How do batch processing and stream processing differ, with representative use cases?
• What does 'document-oriented' refer to in MongoDB?
• What are best practices for ensuring reproducibility and versioning in a data science project?
• Which constraint uniquely identifies a row and cannot be NULL?
• ROW_NUMBER() is typically used for which purpose in SQL queries?
• What is the purpose of a learning curve and how would you interpret it?
• In the bias-variance decomposition, what does the irreducible error represent?
• Besides removing correlated features, which methods can address multicollinearity in regression?
• What is the difference between variance and bias in a machine learning model, and how do regularization techniques affect them?
• In concept drift management, what is a rolling window approach and what is its purpose?
• What are scalar functions?
• Which missingness mechanism describes MCAR and guides simple imputation strategies?
• If you perform a test with p-value 0.051 and alpha 0.05, what is the correct conclusion?
• Which of the following is NOT a scalar function?
• Why does collation matter?
• What is a stored procedure?
• In a query using GROUP BY, how does HAVING operate?
• Which data structure is commonly used to implement database indexes, and what does it store?
• In SQL, what structure enforces the relationship integrity between a parent and its children in a one-to-many relationship?
• What are the advantages of using views?
• MongoDB is best described as which kind of database?
• When would you typically use a Python lambda function?
• What is Auto Increment?
• Data leakage in model evaluation occurs when training data contains information from the test data; which of the following is an example?
• What SQL operation retrieves combined data from related tables?
• In five independent Bernoulli trials with success probability p = 0.3, what is the probability of at least one success?