Data Analyst Interview Questions and Answers (2026 Master Guide)

Q1. What are the key responsibilities of a Data Analyst in an enterprise setting?

Foundation • 0-2 Yrs

A Data Analyst collects raw data from transactional databases/APIs, cleans and models data, performs exploratory data analysis (EDA), identifies trends and anomalies, designs automated BI dashboards, and communicates actionable business recommendations to executive stakeholders.

Q2. What is the difference between Data Analytics and Data Science?

Foundation • 0-2 Yrs

Data Analytics focuses on examining historical datasets to identify patterns, evaluate business performance, and answer 'What happened and why?'. Data Science incorporates advanced statistical modeling, machine learning, and predictive algorithms to build automated models that predict 'What will happen next?'.

Q3. Explain the difference between qualitative and quantitative data.

Foundation • 0-2 Yrs

Quantitative data is numerical, measurable, and objective (e.g., revenue, user age, website visits), analyzed via statistical aggregations. Qualitative data is descriptive, non-numerical, and categorical (e.g., customer survey feedback, review sentiment, product categories).

Q4. What is the difference between structured, semi-structured, and unstructured data?

Foundation • 0-2 Yrs

Structured data resides in fixed relational tables with strict schemas (SQL databases). Semi-structured data contains organizational markers/tags without rigid schemas (JSON, XML, CSV). Unstructured data lacks pre-defined data models (audio files, video, PDFs, unstructured text emails).

Q5. Explain the steps involved in the Data Cleaning and Preprocessing lifecycle.

Foundation • 0-2 Yrs

1. Removing duplicate records -> 2. Handling missing values (imputation via mean/median or deletion) -> 3. Standardizing data types and formats -> 4. Detecting and handling outliers -> 5. Validating constraints and referential integrity -> 6. Normalization/Scaling.

Q6. What are the most common methods to handle missing values in a dataset?

Foundation • 0-2 Yrs

Methods include: Deletion (Listwise/Pairwise deletion when missingness is under 3-5%), Imputation (Mean/Median for skewed numerical data, Mode for categorical data), Advanced Imputation (KNN or regression imputation), or creating a separate 'Unknown' category.

Q7. What is an outlier and how can you detect it statistically?

Foundation • 0-2 Yrs

An outlier is a data point that diverges significantly from the overall distribution. It can be detected using: 1. Z-Score (points beyond ±3 standard deviations from mean in normal distributions); 2. Interquartile Range (IQR = Q3 - Q1, outliers lie below Q1 - 1.5*IQR or above Q3 + 1.5*IQR); 3. Box plots and scatter plots.

Q8. Explain the difference between descriptive, diagnostic, predictive, and prescriptive analytics.

Foundation • 0-2 Yrs

Descriptive: What happened? (Historical KPI reporting). Diagnostic: Why did it happen? (Root-cause drill-downs). Predictive: What is likely to happen? (Statistical forecasting & ML). Prescriptive: What specific action should we take? (Optimization engines and decision trees).

Q9. What is a Pivot Table in Excel and why is it useful?

Foundation • 0-2 Yrs

A Pivot Table is an interactive data summarization tool in spreadsheets that automatically sorts, counts, totals, or averages data stored in large tables without writing complex formulas, facilitating rapid cross-tabulation and executive reporting.

Q10. What is the difference between VLOOKUP and XLOOKUP in Excel?

Foundation • 0-2 Yrs

VLOOKUP searches only from left to right, requires specifying an exact column index number (which breaks when columns are inserted), and defaults to approximate match. XLOOKUP can search in any direction (left or right), uses exact match by default, handles vertical and horizontal lookups, and provides native error handling.

Q11. What are key business metrics: Churn Rate, LTV, CAC, and Retention Rate?

Foundation • 0-2 Yrs

Churn Rate: % of customers who cancel during a given period. CAC (Customer Acquisition Cost): Total marketing & sales spend divided by new customers acquired. LTV (Customer Lifetime Value): Projected net profit generated by a customer over their relationship. Retention Rate: % of existing customers retained over time.

Q12. Explain correlation vs causation with a practical example.

Foundation • 0-2 Yrs

Correlation measures the statistical association between two variables (e.g., ice cream sales and drowning rates increase together). Causation means changes in one variable directly produce changes in the other. A confounding variable (hot weather) drives both, proving that correlation does not imply causation.

Q13. What is an A/B Test and how is it conducted?

Foundation • 0-2 Yrs

An A/B test (split test) is a randomized experiment comparing two versions (A control, B variant). Traffic is split evenly, user interactions are recorded, and statistical hypothesis testing (t-test or chi-square) determines whether the difference in conversion rate is statistically significant.

Q14. What is a P-value and how does it relate to the Significance Level (Alpha)?

Foundation • 0-2 Yrs

The p-value is the probability of observing results at least as extreme as the sample data assuming the Null Hypothesis (H0) is true. If p-value < alpha (typically 0.05), we reject the Null Hypothesis and conclude the observed effect is statistically significant.

Q15. What is the difference between Type I and Type II errors in hypothesis testing?

Foundation • 0-2 Yrs

Type I Error (False Positive, Alpha): Rejecting the Null Hypothesis when it is actually true (e.g., concluding a feature increased sales when it didn't). Type II Error (False Negative, Beta): Failing to reject the Null Hypothesis when it is false (e.g., missing a real positive impact).

Q16. How do you detect and handle data leakage in analytical models?

Experienced • 3-5+ Yrs

Data leakage occurs when information from outside the training dataset (or future information not available at prediction time) is inadvertently included in model training. Prevent it by splitting train/test sets before any normalization, feature scaling, or imputation, and verifying timestamps on features.

Q17. Explain the difference between Pandas loc and iloc with examples.

Experienced • 3-5+ Yrs

loc is label-based indexing: df.loc['row_label', 'column_name'] (inclusive of end bounds). iloc is purely integer-based positional indexing: df.iloc[0:5, 0:2] (exclusive of end bounds, standard 0-indexed Python slicing).

Q18. How do you optimize Pandas performance when handling large datasets (>5 GB)?

Experienced • 3-5+ Yrs

1. Downcast numeric data types (float64 -> float32, int64 -> int16/int8); 2. Convert high-cardinality strings into 'category' dtype; 3. Use chunking (pd.read_csv(chunksize=100000)); 4. Leverage vectorized operations instead of .apply() or iterrows(); 5. Use Polars or DuckDB for multi-threaded querying.

Q19. Explain Cohort Analysis and how it is used to measure user retention.

Experienced • 3-5+ Yrs

Cohort Analysis segments users by a shared acquisition date (e.g., Month of signup). Retention is tracked across subsequent months (Month 1, Month 2, Month 3...) in a triangular retention matrix, revealing whether product onboarding changes improve long-term user stickiness over time.

Q20. How do you identify seasonality vs trends in time series data?

Experienced • 3-5+ Yrs

Time series data can be decomposed into: Trend (long-term upward or downward movement), Seasonality (repeating cyclical fluctuations at fixed calendar intervals like day/month/quarter), and Residuals (noise). Statistical decomposition (statsmodels seasonal_decompose) isolates these components.

Q21. What is the difference between Left Join, Inner Join, and Cross Join in business reporting?

Experienced • 3-5+ Yrs

Left Join retains all primary entities (e.g., all active customers) even if they haven't generated sales (identifying non-buyers). Inner Join retains only intersecting records (customers with verified sales). Cross Join generates a Cartesian product (e.g., combining all stores with all calendar dates to detect missing daily sales logs).

Q22. Explain Simpson's Paradox and how it impacts business decision-making.

Experienced • 3-5+ Yrs

Simpson's Paradox occurs when a statistical trend appears across separate sub-groups of data but disappears or reverses when the groups are aggregated together. It is caused by lurking confounding variables and unequal group sample sizes, emphasizing the need to segment metrics.

Q23. What is the Central Limit Theorem (CLT) and why is it fundamental in analytics?

Experienced • 3-5+ Yrs

The CLT states that the sampling distribution of the sample mean approaches a normal distribution as sample size increases (typically N > 30), regardless of the shape of the underlying population distribution, enabling parametric hypothesis testing on real-world non-normal data.

Q24. How do you calculate and interpret Customer Churn in subscription businesses?

Experienced • 3-5+ Yrs

Customer Churn Rate = (Customers Lost during Period) / (Customers at Start of Period). For net revenue retention, calculate Net MRR Churn = (Churned MRR + Contraction MRR - Expansion MRR) / Starting MRR. Negative net churn indicates expansion revenue exceeds lost revenue.

Q25. Explain the difference between Parametric and Non-Parametric statistical tests.

Experienced • 3-5+ Yrs

Parametric tests (t-test, ANOVA) assume the data follows a specific distribution (usually Normal) with equal variances. Non-parametric tests (Mann-Whitney U, Kruskal-Wallis, Chi-Square) make zero distribution assumptions and are used on skewed data or small sample sizes.

Q26. How do you structure an executive data presentation to non-technical stakeholders?

Experienced • 3-5+ Yrs

Follow the Minto Pyramid Principle: Start with the executive conclusion/recommendation first, followed by supporting grouped arguments, and finish with data evidence. Focus on revenue/cost business impact rather than algorithms, using clean charts with highlighted callouts.

Q27. What is the difference between Dimension tables and Fact tables?

Experienced • 3-5+ Yrs

Fact tables contain measurable numeric metrics recorded from business events (e.g., SalesAmount, Quantity, DiscountID, timestamps). Dimension tables contain descriptive attributes that provide context to facts (e.g., CustomerName, StoreCity, ProductCategory, DateKey).

Q28. How do you handle severe class imbalance in analytical datasets?

Experienced • 3-5+ Yrs

Techniques include: 1. Resampling (undersampling majority class or SMOTE oversampling minority class); 2. Using appropriate evaluation metrics (ROC-AUC, Precision-Recall curve instead of misleading accuracy); 3. Adjusting classification probability thresholds.

Q29. Explain RFM (Recency, Frequency, Monetary) Segmentation.

Experienced • 3-5+ Yrs

RFM segments customers based on: Recency (how recently did they purchase?), Frequency (how often do they purchase?), and Monetary value (how much do they spend?). Scoring users 1-5 across each dimension groups them into Champions, Loyalists, At-Risk, or Churned tiers.

Q30. What is Data Governance and why is Data Lineage critical in analytics?

Experienced • 3-5+ Yrs

Data Governance establishes data quality, privacy (GDPR/HIPAA), and security policies across an enterprise. Data Lineage maps the lifecycle and path of data from origin source systems through ETL pipelines to final BI reports, ensuring auditability and impact analysis when schemas change.

Q31. Write a SQL query to find the 2nd highest salary from an Employee table without using LIMIT.

Live Coding Challenge

We use a subquery with MAX() filtering salaries below the overall maximum salary.

SELECT MAX(salary) AS SecondHighestSalary
FROM Employee
WHERE salary < (
    SELECT MAX(salary) FROM Employee
);

Q32. Write a Python Pandas script to identify and remove duplicate rows while keeping the latest entry.

Live Coding Challenge

We use sort_values on the timestamp column and drop_duplicates keeping the last entry.

import pandas as pd

def clean_duplicates(df):
    df_sorted = df.sort_values(by='updated_at')
    return df_sorted.drop_duplicates(subset=['customer_id'], keep='last')

Q33. Write a SQL query using window functions to calculate Month-over-Month Revenue growth.

Live Coding Challenge

We aggregate monthly sales, use LAG() to retrieve previous month revenue, and compute percentage difference.

WITH MonthlySales AS (
    SELECT 
        DATE_TRUNC('month', order_date) AS sales_month,
        SUM(amount) AS current_revenue
    FROM orders
    GROUP BY 1
)
SELECT 
    sales_month,
    current_revenue,
    LAG(current_revenue) OVER (ORDER BY sales_month) AS prior_revenue,
    ROUND(100.0 * (current_revenue - LAG(current_revenue) OVER (ORDER BY sales_month)) 
          / LAG(current_revenue) OVER (ORDER BY sales_month), 2) AS mom_growth_pct
FROM MonthlySales;

Q34. Write a Python script to detect outliers in a Pandas Series using the IQR method.

Live Coding Challenge

We calculate Q1 and Q3, compute IQR, and filter points outside 1.5 * IQR bounds.

import numpy as np
import pandas as pd

def detect_outliers_iqr(series: pd.Series):
    q1 = series.quantile(0.25)
    q3 = series.quantile(0.75)
    iqr = q3 - q1
    lower_bound = q1 - 1.5 * iqr
    upper_bound = q3 + 1.5 * iqr
    return series[(series < lower_bound) | (series > upper_bound)]

Q35. Write a SQL query to find the top 3 spending customers in each city.

Live Coding Challenge

We group by city and customer, calculate total spend, rank customers using DENSE_RANK(), and filter rank <= 3.

WITH RankedCustomers AS (
    SELECT 
        city,
        customer_name,
        SUM(order_amount) AS total_spend,
        DENSE_RANK() OVER(PARTITION BY city ORDER BY SUM(order_amount) DESC) as rnk
    FROM customer_orders
    GROUP BY city, customer_name
)
SELECT city, customer_name, total_spend
FROM RankedCustomers
WHERE rnk <= 3;

Q36. Write a Python function to impute missing values with the median of each group.

Live Coding Challenge

We use transform with median inside a groupby expression.

import pandas as pd

def impute_by_group(df, group_col, value_col):
    df[value_col] = df.groupby(group_col)[value_col].transform(
        lambda x: x.fillna(x.median())
    )
    return df

Q37. Write a SQL query to calculate the rolling 7-day average sales.

Live Coding Challenge

We use AVG() OVER with an ORDER BY date and ROWS BETWEEN 6 PRECEDING AND CURRENT ROW frame.

SELECT 
    order_date,
    daily_sales,
    AVG(daily_sales) OVER (
        ORDER BY order_date
        ROWS BETWEEN 6 PRECEDING AND CURRENT ROW
    ) AS rolling_7day_avg
FROM daily_sales_summary;

Q38. Write a Python script to perform an Independent Two-Sample T-Test.

Live Coding Challenge

We use scipy.stats.ttest_ind to compare means between variant A and variant B.

from scipy import stats

def run_ab_test(control_group, variant_group):
    t_stat, p_val = stats.ttest_ind(control_group, variant_group, equal_var=False)
    is_significant = p_val < 0.05
    return {"t_statistic": t_stat, "p_value": p_val, "significant": is_significant}

Q39. Write a SQL query to calculate user retention rate 30 days after signup.

Live Coding Challenge

We self-join users with activity logs between day 1 and day 30.

SELECT 
    DATE_TRUNC('month', u.signup_date) AS cohort_month,
    COUNT(DISTINCT u.user_id) AS total_signups,
    COUNT(DISTINCT a.user_id) AS retained_users,
    ROUND(100.0 * COUNT(DISTINCT a.user_id) / COUNT(DISTINCT u.user_id), 2) AS retention_pct
FROM users u
LEFT JOIN user_activity a 
    ON u.user_id = a.user_id 
    AND a.activity_date BETWEEN u.signup_date AND u.signup_date + INTERVAL '30 days'
GROUP BY 1;

Q40. Write a Python Pandas one-liner to pivot a transaction table into a monthly customer matrix.

Live Coding Challenge

We invoke pivot_table aggregating sum of amounts with 0 filled for missing periods.

pivot_df = df.pivot_table(index='customer_id', columns='order_month', values='amount', aggfunc='sum', fill_value=0)