Model answers for 4 data analyst questions
Typical rounds: SQL test, technical interview, case or presentation
Q: Daily orders dropped 15% yesterday. How would you investigate? A: First confirm it's real: check data pipeline delays and tracking changes. Then segment the drop by platform, city, new vs returning users and funnel step to find where it is concentrated, compare with the same weekday last week, and check for releases, outages, holidays or competitor promotions.
Q: How do you handle missing data? A: Find out why it's missing first. Then drop rows if the share is small and random, impute with median or a business default if reasonable, or flag it as a separate category. I always note the choice and its effect in the report.
Q: What is the difference between mean and median, and when do you use each? A: The mean is the average and is pulled by outliers; the median is the middle value. For skewed data like order value or salary, the median gives a more honest typical value.
Q: Tell me about a dashboard you built. A: I built a Power BI sales dashboard for 12 regional managers with filters by region and product. It replaced a weekly Excel report, saved about 6 hours a week, and helped managers spot two slow product lines early.