2–5 years · Mid — SQL · Excel · Power BI / Tableau · Statistics · Python · Stakeholder reporting
Calculates the percentage of missing values per column; examines patterns of missingness to identify randomness; evaluates the impact on analyses; considers imputation methods or dropping rows/columns; documents assumptions made in handling missing data.
What if the missing values are not random and you suspect bias?
Defines null and alternative hypotheses; selects the appropriate statistical test, such as a t-test or chi-square test; checks assumptions for normality and variance; calculates p-values; interprets results in the context of confidence intervals.
What actions do you take if the sample size is too small to draw conclusions?
Revisits the data source to ensure accuracy; checks the DAX calculations for errors; validates the aggregation methods used in visuals; compares results with raw data; engages with the stakeholders to clarify their expectations and definitions.
What if the stakeholders insist on their interpretation despite your findings?
Defines the key questions stakeholders want answered; selects the most relevant subset of data; employs summary statistics and visualisations for quick insights; communicates assumptions and limitations clearly; schedules a follow-up for deeper analysis post-deadline.
After delivering the initial insights, stakeholders want more detailed analysis. How do you manage their expectations?
Names the specific migration context and data involved; details the unexpected issues faced and their impact on the project timeline; explains the steps taken to rectify the issue and how it affected the final outcome; reflects on lessons learned regarding planning and testing.
What would you do differently in future migrations?
Describes the specific feature and its intended impact; identifies the key assumptions made during the design phase that proved incorrect; mentions the metrics used to evaluate success and how they fell short; discusses any changes made after the failure and their outcomes.
In hindsight, what additional data would have informed your decision?
Load six more — you will get questions you have not seen first.
Give me six more →Say your answer out loud before revealing. Reading the answer first feels like learning and is not.