Data Analyst Interview Prep India 2026

"Your company's return rate jumped from 8% to 14% over the last quarter. The head of operations wants to understand why and what to do about it."

Orders table: order_id, product_id, category, city, order_date, amount, payment_method
Returns table: return_id, order_id, return_date, return_reason, refund_amount

1. Define metric: return rate = returns / orders, by week to see trend start
2. Segment: by category, city, payment method, product — where is the spike concentrated?
3. Look at return_reason distribution — is it product quality, delivery damage, or wrong item?
4. Compare return rate for new vs. repeat customers
5. Check if specific products or sellers are driving disproportionate returns

"Build a Power BI dashboard for the Regional Sales Manager who oversees 8 states and 200+ distributors. They currently use a 40-tab Excel file updated manually every Monday."

Sales data: date, distributor_id, state, product_sku, units_sold, revenue, target
Distributor master: distributor_id, name, state, region, tier

1. Identify the 3 questions the manager asks every Monday: total revenue vs. target, top/bottom 5 distributors, state-wise breakdown
2. Design 3-4 pages: Executive Summary → State Drill-down → Distributor Performance → SKU Analysis
3. Add slicers for state, date range, and distributor tier
4. Use conditional formatting (red/amber/green) for target achievement
5. Replace the 40-tab Excel with one dashboard that auto-refreshes from a single data source

"The finance team spends 3 hours every Monday morning manually copying data from 5 Excel files into a master sheet and then creating 8 standard reports. Automate this."

5 regional Excel files (same structure): date, branch_code, product_type, disbursement_amount, collection_amount, outstanding
Master template: summary sheet + 8 pre-built pivot report tabs

1. Map all 5 input files — same columns? Any formatting inconsistencies?
2. Write Python (Pandas + openpyxl) script to read all 5, concatenate, validate, and write to master
3. Automate the 8 pivot reports using openpyxl or xlwings
4. Schedule with Windows Task Scheduler (or a Monday-morning email trigger)
5. Document the script so the team can maintain it

I was building a weekly revenue report for the sales team at my internship. The numbers for one region looked unusually high — about 40% above the previous week with no business reason.

My task was to investigate before the report went to the VP of Sales.

I traced the data back to its source — the raw transaction table. I found that a data pipeline had double-counted transactions for that region because of a duplicate key error in the ETL job. I documented the issue, fixed the query by adding a DISTINCT clause, and recalculated the affected 3 weeks of reports. I also added a row-count validation check to the ETL process to catch this in future.

The corrected report went out on time. The VP caught it in review and thanked the team for the correction. The validation check has since caught two more pipeline issues before they reached reporting.

In my last project, I built a customer segmentation analysis using RFM (Recency, Frequency, Monetary) scoring. The output was a matrix of 27 segments. The marketing head — who had no analytics background — needed to understand it to allocate campaign budgets.

I had to translate a technically correct but unusable output into something actionable within a 20-minute meeting.

I collapsed the 27 segments into 4 plain-language groups: Champions (buy often, recently, high value), At Risk (used to buy a lot, not recently), New (bought once, recently), and Dormant (not bought in 6+ months). For each group, I prepared one slide: how many customers, what percentage of revenue, and one recommended campaign action with an estimated reach.

The marketing head allocated budgets within the meeting. The "At Risk" re-engagement campaign generated a 22% reactivation rate in the first month — the highest the team had seen.

The marketing team had run a 3-week influencer campaign and was expecting to show strong ROI in the monthly review. Before the meeting, I analysed the campaign data and found that conversion rate from influencer traffic was 0.4% vs. 2.1% from organic search.

I had to present findings that contradicted the team's expectation — in a review meeting they had already prepared slides for.

I shared the finding with the marketing manager 24 hours before the meeting — not in the meeting itself. I framed it as "here is something we should look at together before the presentation." We reviewed the data together, verified there were no tracking errors, and agreed to present the full picture honestly. I also found a positive finding: the influencer campaign had driven 3× the usual branded search volume the following week — delayed conversion.

The presentation went well. The team proposed a revised influencer strategy with longer attribution windows. I was asked to build a dashboard tracking branded search lift after future campaigns.

3-5 bullet points. Each bullet: one insight + one supporting number + one implication. This is what the hiring manager reads first — if it is unclear, the rest does not matter.

Show you cleaned the data: row count before and after removing duplicates/nulls, date range, key columns and data types. This proves analytical rigour.

3-5 insights maximum. Each on its own slide/section: insight headline → supporting chart → 1-2 sentence explanation. Never bury an insight in a paragraph.

1-2 specific actions. Use the format: "I recommend X because the data shows Y, which I estimate will lead to Z." Avoid vague recommendations like "improve marketing".

Your SQL queries, Python code, or Power BI data model. Clearly commented. Shows your working — interviewers sometimes review code separately.

When asked "what is your expected CTC?", respond: "I am open based on the total package — what is the range budgeted for this role?"

Variable pay, joining bonus, learning budget, and WFH flexibility all have monetary value. A ₹7L offer with ₹50K learning budget beats a ₹7.2L offer without one.

Standard in India. Companies expect a counter. Say: "Based on my research, could we look at [X]? I am excited about the role and this would help me make the decision quickly."

If you have another offer, you can mention the range — not the company. "I have another offer in the ₹X-Y range — is there flexibility here?" Only say this if it is true.