Data Science Interview Questions and Answers
By Karthick Selvam · Published
Frequently asked data science interview questions.
What is a p-value?
The probability of observing results at least as extreme as yours if the null hypothesis were true. It is not the probability that the hypothesis is true.
How do you design an A/B test?
Define a metric and hypothesis, calculate sample size, randomise, run for a fixed period and analyse with an appropriate test.
What is cross-validation?
Splitting data into folds to train and validate repeatedly, giving a more reliable estimate of performance.
What is data leakage?
When information unavailable at prediction time influences training, inflating performance.
Explain regularisation.
Adding a penalty for model complexity (L1, L2) to reduce overfitting.
How would you explain a model to a business user?
Focus on the decision, key drivers and uncertainty, using simple visuals.
Related programmes
- Advanced SQL for Data Professionals
- Power BI for Business Analytics
- Python for Data Engineering & Analysis
Turn this into a team capability
Our corporate programmes cover topics like this hands-on, mapped to your stack.
Data & Analytics guides
Python vs SQL for Data Work: Which to Learn First?
SQL and Python solve different parts of data work. Most data professionals need both.
What Is Power BI?
Power BI is Microsoft's business intelligence platform for connecting to data, modelling it and building interactive reports.
What Is Data Science?
Data science combines statistics, programming and domain knowledge to extract insights and build predictive models from data.