Data Analyst Interview Questions India 2026: SQL, Python and Business Case Guide

SQL is the primary technical filter in Indian data analyst interviews. The questions are real-world analytical: you are given a table schema (orders, users, products, sessions) and asked to extract business insights, not just test your knowledge of syntax. High-priority SQL topics for 2026: Window Functions: these appear in 80% of advanced data analyst SQL rounds. RANK() vs DENSERANK() vs ROWNUMBER() (the difference matters when there are ties). LAG() and LEAD() for comparing a metric to the previous or next period. SUM() OVER (PARTITION BY userid ORDER BY orderdate) for running totals. Running 7-day average sales by store using window functions. Common Table Expressions (CTEs): used for multi-step queries. 'Find the users who placed their first order in Q4 and placed at least 2 more orders in Q1 of the following year.' This requires a CTE to identify first orders, then join back to find repeat orders. Self Joins: 'Find all customers who placed an order this month but did not place an order in the same month last year.' Self joins on the orders table with date filtering. Subqueries and IN vs EXISTS: when to use each and the performance difference on large tables. Date functions: DATEDIFF, DATETRUNC, DATEFORMAT, EXTRACT (month, year, weekday) in MySQL, PostgreSQL, and BigQuery syntax (know which your target company uses). Handling NULL: COALESCE, NULLIF, and why NULL comparisons require IS NULL rather than = NULL.

Python interviews for data analyst roles in India test practical data manipulation skills, not academic computer science. Core pandas operations that appear in coding assessments: readcsv with handling of common issues (mixed dtypes, missing headers, encoding issues). groupby with multiple aggregation functions: df.groupby('city').agg({'orders': 'sum', 'revenue': 'mean', 'users': 'nunique'}). merge (inner, left, right, outer) and when each is appropriate: almost identical to SQL JOINs. pivottable for reshaping data: sales by product category and month. apply with lambda for custom column transformations. Handling missing values: isnull().sum() to audit, fillna() with mean/median/mode/forward-fill depending on context, dropna() with subset and thresh parameters. Datetime operations: pd.todatetime(), .dt.month, .dt.dayofweek, .dt.quarter. String operations with .str accessor: .str.contains(), .str.extract(), .str.split(). mergeasof for time-series merges (matching transactions to the most recent price at the time of the transaction). Common interview format: you are given a CSV or a SQL database dump and asked to write a 30-minute EDA (exploratory data analysis) in a Jupyter notebook that produces 3-5 business insights with supporting charts. Matplotlib and seaborn: bar chart, line chart, scatter plot, histogram, box plot. Know how to label axes, set titles, and format tick labels.

The business case component tests whether you can think like an analyst, not just compute like one. Common formats at Indian companies: Metric investigation: 'Swiggy's daily active users dropped 12% last Tuesday in Mumbai. How would you investigate?' Expected approach: clarify the metric definition (DAU: app opens, or orders?), segment by possible dimensions (platform iOS vs Android, user cohort, city area), identify correlated events (competitor promotion, local holiday, infrastructure incident, app update release), quantify each hypothesis with data, and recommend a next action. Dashboard design: 'Design a KPI dashboard for a quick commerce store manager to track their 10 outlets in Bengaluru.' Expected approach: identify the store manager's decisions (staffing, inventory reorder, pricing changes), map each decision to a leading metric (orders per hour by outlet, pick time, out-of-stock rate, basket size), structure the dashboard from summary to detail (top-level scorecard, then drilldowns). A/B test interpretation: 'We ran an experiment: 50% of users saw a redesigned checkout, 50% saw the original. The redesign has 4% higher add-to-cart rate but 3% lower order completion rate. What do you recommend?' Expected approach: check statistical significance of both metrics, identify the funnel step where the drop happens, hypothesise why higher add-to-cart leads to lower completion (friction at payment step?), and recommend a segmented analysis before a full rollout decision.

Data analyst compensation depends heavily on the company type and city. 2026 benchmarks: Junior data analyst (0-2 years): Rs 4-10 LPA at traditional companies (banks, FMCG, manufacturing), Rs 8-18 LPA at product companies (Swiggy, Flipkart, Ola, Paytm) and consulting firms. Mid-level data analyst (2-5 years): Rs 10-25 LPA at traditional companies, Rs 20-45 LPA at product companies. Senior data analyst / Analytics Manager (5+ years): Rs 20-45 LPA at traditional companies, Rs 40-80 LPA at product companies and startups. Bengaluru and Mumbai pay 20-30% above equivalent roles in Hyderabad, Pune, or Chennai. Career paths from data analyst: Data Scientist (add ML skills, Python modelling libraries, statistics): Rs 30-80 LPA. Analytics Engineer (add dbt, data pipeline skills, data warehouse): Rs 25-60 LPA. Business Intelligence Analyst (deeper BI tools: Tableau, Looker, Power BI): Rs 15-40 LPA. Product Analyst / Growth Analyst (product-focused track with A/B testing, retention, conversion): Rs 20-55 LPA. The transition to data scientist requires: statistics depth (probability, hypothesis testing, regression), Python ML libraries (scikit-learn, XGBoost), and ideally a portfolio project showing end-to-end ML model development.