Top 50 Data Analyst Interview Questions for Freshers Asked in Top MNCs

1. What does a Data Analyst do?

A Data Analyst collects, cleans, analyzes, and interprets data to help a business make informed decisions.
A typical analytics workflow includes understanding the business requirement, extracting data, cleaning and validating it, performing analysis, and communicating actionable findings.

2. What is the typical data analysis process?

A basic data analysis process can be explained as:

1. Understand the business problem.
2. Identify the required data.
3. Extract and clean the data.
4. Validate data quality.
5. Perform exploratory analysis.
6. Identify patterns and insights.
7. Communicate the findings.
8. Recommend appropriate actions.

3. How would you approach a completely new dataset?

First, understand the business context and the purpose of the dataset.
Then examine its schema, data types, missing values, duplicates, abnormal values, inconsistent formats, and relationships between columns.
Only after validating the data would you begin deeper analysis.

4. How do you handle missing values?

There is no single solution for every dataset.
Depending on the situation, you may:

• Remove records when the missing portion is negligible.
• Replace values using an appropriate statistical measure.
• Use interpolation for suitable time-series data.
• Use predictive techniques when justified.

The decision should depend on why the data is missing and how the missing values affect the analysis.

5. How would you explain a technical finding to a non-technical stakeholder?

Avoid unnecessary technical terminology and connect the finding to a business outcome.
For example, instead of explaining a complex statistical calculation, explain how the finding affects revenue, customer retention, operational efficiency, or another relevant business metric.

6. What is the difference between INNER JOIN and LEFT JOIN?

INNER JOIN returns records that have matching values in both tables.
LEFT JOIN returns every record from the left table and matching records from the right table. If there is no match, the right-side columns contain NULL.
A LEFT JOIN is particularly useful when you want to identify records without a corresponding match.

7. What is an anti-join?

An anti-join identifies records in one table that do not have a corresponding record in another table.
For example, finding customers who registered but never placed an order can be achieved using a LEFT JOIN with a NULL condition or NOT EXISTS.

8. What is a CROSS JOIN?

A CROSS JOIN produces the Cartesian product of two tables.
If one table contains 10 rows and another contains 5 rows, the result can contain 50 combinations.
It is useful when every possible combination is intentionally required, but it can become computationally expensive.

9. What is the difference between UNION and UNION ALL?

UNION combines result sets and removes duplicate records.
UNION ALL combines result sets while retaining duplicates.
Because UNION ALL does not perform deduplication, it can be more efficient when duplicate removal is not required.

10. What is a window function?

A window function performs a calculation across related rows while preserving the individual rows in the result.
Common examples include:

• ROW_NUMBER()
• RANK()
• DENSE_RANK()
• LAG()
• LEAD()

They are commonly used for rankings, running totals, comparisons, and time-based analysis.

21. What is the difference between a Series and a DataFrame in Pandas?

A Series is a one-dimensional labeled data structure.
A DataFrame is a two-dimensional data structure consisting of rows and columns.
A DataFrame can contain multiple Series representing different columns.

22. What is the difference between loc and iloc?

.loc performs label-based selection.
.iloc performs integer-position-based selection.
For example, .loc can select a row based on its index label, while .iloc selects based on its numerical position.

23. How do you handle missing values in Pandas?

Common methods include:

• dropna() to remove missing records.
• fillna() to replace missing values.
• interpolate() for suitable datasets, particularly time-series data.

The appropriate method depends on the analytical context.

24. What is vectorization in Pandas?

Vectorization means performing operations on entire columns or arrays rather than processing rows individually with explicit loops.
It is generally more efficient for large datasets.

25. Why can iterrows() be inefficient?

iterrows() processes a DataFrame row by row.
For large datasets, this can be considerably less efficient than vectorized operations, Boolean masking, or other column-based approaches.

26. How would you process a CSV file larger than your computer's RAM?

One approach is to process the file in chunks using Pandas' chunksize parameter.
You can also:

• Load only required columns.
• Optimize data types.
• Convert appropriate columns to categorical types.
• Use columnar formats such as Parquet.
• Consider analytical engines such as DuckDB or Polars for larger-than-memory workloads.

27. What is the benefit of converting a dataset to Parquet?

Parquet is a columnar storage format that can reduce storage requirements and allow efficient column-level access.
This can be useful when only a subset of columns is required for analysis.

28. What is the difference between VLOOKUP and XLOOKUP?

VLOOKUP searches for a value in the first column of a selected range and generally returns a value from a column to its right.
XLOOKUP separates the lookup array from the return array, allowing more flexible lookups and making it less dependent on column position.

29. What is a Pivot Table?

A Pivot Table summarizes large datasets by grouping and aggregating information.
It can quickly show metrics such as sales by region, revenue by product, or transactions by month.

30. What are SUMIFS and COUNTIFS used for?

SUMIFS calculates a sum based on multiple conditions.
COUNTIFS counts records that meet multiple conditions.
They are useful for conditional reporting and analysis in Excel.

31. How would you clean messy data in Excel?

You can use techniques such as:

• Removing duplicates.
• Trimming unnecessary spaces.
• Standardizing formats.
• Applying data validation.
• Using text functions.
• Using Power Query for repeatable cleaning workflows.

32. What is the difference between a Power BI measure and a calculated column?

The research recommends measures for dynamic analytical calculations such as percentages and KPI calculations.

33. What is DAX?

DAX stands for Data Analysis Expressions.
It is the formula language used in Power BI for creating calculations and measures.

34. What is the difference between row context and filter context?

Row context evaluates an expression in relation to an individual row.
Filter context represents the filters affecting a calculation, such as slicers, visual selections, and report filters.
Understanding both is important for writing effective DAX calculations.

35. What does CALCULATE() do in DAX?

CALCULATE() evaluates an expression after modifying the existing filter context.
It is one of the most important DAX functions for creating dynamic calculations.

36. What is context transition?

Context transition occurs when CALCULATE() converts an existing row context into an equivalent filter context.
This is an important concept when working with calculated columns and iterator functions.

37. What is the difference between Import Mode and DirectQuery?

Import Mode loads data into Power BI's in-memory engine.
DirectQuery leaves the data at the source and sends queries back to that source when data is requested.
Import generally provides fast report interaction, while DirectQuery can be useful when data needs to remain at the source or near-real-time access is required.

38. What are Tableau LOD expressions?

Level of Detail expressions allow calculations at a specified level of granularity independently of the visualization's displayed level of detail.
The three primary types are:

• FIXED
• INCLUDE
• EXCLUDE

39. What is an A/B test?

An A/B test compares two variants, such as a control and treatment group, to evaluate whether a change produces a measurable difference.
The purpose is to make a controlled comparison rather than relying only on observational differences.

40. What factors determine the required duration of an A/B test?

Test duration depends on factors including:

• Required sample size.
• Significance level.
• Statistical power.
• Minimum Detectable Effect (MDE).
• Expected traffic.
• Business cycles and behavioral variation.

The research also highlights the importance of capturing relevant weekday/weekend variation rather than stopping immediately after the mathematical sample requirement is reached.

41. What is the peeking problem in A/B testing?

Peeking occurs when analysts repeatedly check experiment results and stop the test as soon as statistical significance appears.
Under a standard fixed-sample frequentist design, repeated unplanned checking can increase the probability of false positives.

42. What is Sample Ratio Mismatch?

Sample Ratio Mismatch, or SRM, occurs when the observed allocation between experiment groups differs materially from the planned allocation.
For example, if an experiment is designed for a 50/50 split but receives a substantially different distribution, the analyst should investigate the underlying randomization or tracking process before interpreting the experiment results.

43. What is the difference between Frequentist and Bayesian testing?

A Frequentist approach generally evaluates evidence using concepts such as fixed sample sizes and p-values.
A Bayesian approach incorporates prior information and updates beliefs using observed data.
The two approaches differ in how uncertainty and evidence are represented and interpreted.

44. Daily Active Users dropped by 15%. How would you investigate?

Do not immediately assume that users have actually disappeared.
Use a structured investigation:

• Validate the metric and data pipeline.
• Segment the metric by platform, geography, cohort, and product area.
• Identify when the decline started.
• Examine the underlying funnel components.
• Form and prioritize testable hypotheses.

This framework helps distinguish a genuine business change from a tracking or data-quality problem.

45. How would you investigate a sudden drop in sales?

Start by confirming that the decline is real.
Then segment sales by dimensions such as:

• Product
• Region
• Customer type
• Channel
• Time period

Next, compare the timing with relevant business events and examine the components contributing to total sales.

46. What would you do if your analysis contradicts a stakeholder's assumption?

First, verify the data, calculations, and methodology.
If the result remains valid, explain the finding objectively using business-focused language and supporting evidence.
The goal should be to help the stakeholder understand the evidence rather than simply proving them wrong.

47. How would you prioritize different possible causes of a metric decline?

First narrow the problem through segmentation and validation.
Then prioritize hypotheses based on factors such as:

• Likelihood.
• Business impact.
• Ease of testing.
• Evidence already available.

This avoids spending time investigating highly speculative explanations before more testable causes.

48. What is PII and how should an analyst handle it?

PII, or Personally Identifiable Information, refers to information that can identify an individual.
Analysts should handle sensitive information carefully through appropriate controls such as data masking, anonymization, pseudonymization, access controls, and data minimization.
The research also highlights GDPR and CCPA as examples of privacy frameworks analysts may encounter.

49. What should a good Data Analyst portfolio project contain?

A strong project should begin with a clear business question rather than simply demonstrating a tool.
It should show:

• Business objective.
• Data cleaning.
• Analytical methodology.
• Relevant calculations.
• Visualizations.
• Key findings.
• Actionable recommendations.
• Clear documentation.

The research emphasizes depth and realistic problem-solving over simply having a large number of superficial projects.

50. If AI generates your SQL or Python code during an interview, how would you validate it?

Do not blindly accept the generated output.
Review the logic, check assumptions, test edge cases, verify the result against the business requirement, and explain the code to the interviewer.
The research identifies this ability to review, debug, validate, and explain AI-generated work as increasingly relevant to modern analytics interviews.