Prepare for a Data Internship Interview: Tasks and Examples

An internship interview usually gives you limited time to show how you learn, reason, and explain your own work. Prepare evidence at the level you actually have: coursework, a personal project, volunteer work, or a previous placement. Do not invent production impact to make a student project sound larger.

Translate the role into three practice tasks

Read the current description and identify the work it actually names. For a data internship, that might be writing a join, investigating a metric change, and presenting a project. Ask the recruiter about format, permitted tools, duration, and accessibility arrangements if needed. A generic “data science” title does not tell you whether the assessment emphasizes SQL, statistics, software, or communication.

Choose one practice task for each relevant skill. Prepare a small reproducible dataset and an explanation of the result. Familiarity with a library name is less useful than being able to check a simple output.

Prepare a truthful project story

Use this structure: the question, the available data, your contribution, an important choice, how you checked the work, the result, and a limitation. Keep “I” for your contribution and “we” for team work. If you did not measure an outcome, say what you produced and how you validated it.

A fictional example is: “For a class project, I compared delivery times across three service zones. I wrote the cleaning and summary code. I found missing timestamps and reported their frequency before comparing medians. The project did not establish why zones differed, because assignment and distance were not controlled.” The useful feature is its specificity and limits, not a claim of impressive commercial results.

Work through a small analysis task

Question: calculate the mean of observed delivery times and report data completeness. A missing duration is not a zero-minute delivery. Here is a complete Python example using synthetic data:

from statistics import mean

durations = [12, None, 18, 0, 30]
observed = [value for value in durations if value is not None]
average = mean(observed) if observed else None
print({"rows": len(durations), "observed": len(observed),
       "missing": len(durations) - len(observed), "mean": average})

The result has five rows, four observed values, one missing value, and a mean of 15. Zero is retained because it is an observed value in this exercise. In real delivery data, investigate whether zero is valid before including it; the snippet does not establish a domain policy. If every value is missing, the mean is None rather than zero.

Explain why a filter such as if value would be wrong here: it would remove the observed zero. Then state what you would check next, such as negative values, units, duplicate deliveries, and whether missingness differs by zone. The supplied example assumes all nonmissing values are already valid numeric durations.

Communicate while solving

Restate the task and clarify its important ambiguity. State your approach, work through a tiny example, and verify a boundary case. If you get stuck, explain the specific obstacle and try a smaller case. If the interviewer provides a correction, incorporate it and describe how it changes the answer.

For behavioral questions, use one concrete event. Explain what you did when a deadline, disagreement, or unfamiliar concept arose, and what you would change next time. Avoid claiming you never make mistakes or giving a polished story that conceals your own role.

Run a 45-minute preparation session

1. Spend five minutes reviewing the role and selecting a representative task.
2. Spend ten minutes delivering and questioning your project story.
3. Spend twenty minutes solving a technical task without copying an answer.
4. Spend five minutes checking an edge case and explaining a limitation.
5. Spend five minutes recording one observed weakness and a specific follow-up exercise.

Use a structured mock interview for feedback. Track observable behaviors, such as forgetting the denominator, rather than a vague judgment that you “lack confidence.”

Ask about learning and supervision

Useful questions include: What could an intern own by the end of the placement? Who reviews the work? How often is feedback given? What data and tools can interns access? What would a successful project handoff contain? These answers help you understand the opportunity; a famous employer name does not substitute for a workable learning arrangement.

Finish preparation by checking interview logistics and keeping your resume consistent with the projects you can explain. Preparation improves your ability to demonstrate your skills; it cannot guarantee an offer.