Data Analyst Skills in 2026: What to Learn, How Each Skill Is Tested in Interviews, and How to Prove It in a Portfolio

What skills does a data analyst need in 2026?

Most analyst job posts in India combine a short list of tools with a longer list of softer expectations such as "strong problem solving" or "stakeholder management". The table below turns both into specific skills, with how each is typically tested and what evidence to show.

You do not need all of these at expert level. A common fresher profile is strong SQL and Excel, one BI tool, working Python, and clear communication. Our guide to becoming a data analyst in India covers the order to learn them in.

SQL: the skill almost every analyst interview tests

SQL is how analysts get data out of company databases, so it is usually the first technical filter. Expect a timed test on an online platform, or a live round where you share your screen and talk through a query.

A typical question: "For each city, what share of customers ordered more than once in June?" Here is a small sample table you can create in PostgreSQL, MySQL 8 or SQLite:

CREATE TABLE orders (order_id INT, customer_id TEXT, city TEXT, order_date DATE, amount INT);
INSERT INTO orders VALUES
(1,'A','Pune','2026-06-02',1200),(2,'B','Pune','2026-06-05',800),(3,'A','Pune','2026-06-20',1500),
(4,'C','Jaipur','2026-06-03',600),(5,'D','Jaipur','2026-06-11',2200),(6,'E','Jaipur','2026-06-15',900),
(7,'D','Jaipur','2026-06-28',1100),(8,'F','Kochi','2026-06-09',3000),(9,'G','Kochi','2026-06-19',700);

A clean answer uses a CTE:

WITH customer_orders AS (
  SELECT city, customer_id, COUNT(*) AS n_orders
  FROM orders
  WHERE order_date >= '2026-06-01' AND order_date < '2026-07-01'
  GROUP BY city, customer_id
)
SELECT city,
       COUNT(*) AS customers,
       SUM(CASE WHEN n_orders > 1 THEN 1 ELSE 0 END) AS repeat_customers,
       ROUND(100.0 * SUM(CASE WHEN n_orders > 1 THEN 1 ELSE 0 END) / COUNT(*), 1) AS repeat_rate_pct
FROM customer_orders
GROUP BY city
ORDER BY repeat_rate_pct DESC;

On those nine orders, it returns:

city    customers  repeat_customers  repeat_rate_pct
Pune    2          1                 50.0
Jaipur  3          1                 33.3
Kochi   2          0                 0.0

What the interviewer is really checking: do you group at the right level first (customer, then city), and do you notice that two or three customers per city is far too few to call Pune "better"? Saying that out loud earns as much credit as the query. Our SQL interview questions guide has more practice.

Portfolio proof: one project where SQL does the heavy lifting, such as a cohort retention table or a funnel, with the queries saved and commented.

Excel and Google Sheets: still tested, often as a take-home

Many analyst teams still share work in spreadsheets, and plenty of first analyst tasks involve cleaning an export and producing a summary. Interviewers often send a workbook and ask for a summary tab within a time limit.

Know these well: PivotTables, XLOOKUP (or INDEX/MATCH), SUMIFS and COUNTIFS, IF with AND/OR, text functions such as TRIM, date functions, and conditional formatting. For example, June revenue for Pune from a table with city in column B, date in column C and amount in column D:

=SUMIFS(D:D, B:B, "Pune", C:C, ">="&DATE(2026,6,1), C:C, "<"&DATE(2026,7,1))

Portfolio proof: a workbook with raw data, a cleaning tab that shows your steps, and a one-screen summary. Our advanced Excel guide lists the skills analysts are tested on.

BI tools: Power BI, Tableau or Looker Studio

Employers want to know you can turn data into a dashboard someone will actually use. The tool matters less than you think, as long as it matches the job posts you are targeting. Our Power BI vs Tableau comparison helps you pick one.

BI skills are rarely tested with a live build. More often you are asked to walk through a dashboard you made: why these metrics, why this chart type, what a manager should do after looking at it. If you use Power BI, expect questions on measures and filter context; our guide to DAX functions in Power BI covers the common ones.

Portfolio proof: one dashboard with a clear audience ("for a regional sales manager"), 4–6 visuals, and a note of the three things it shows.

Statistics: enough to avoid wrong conclusions

Analyst roles rarely need advanced maths, but they do need you to avoid confident mistakes. Interviewers test this with short conceptual and case questions rather than formulas.

Know these: mean vs median and when each misleads, spread (standard deviation, percentiles), distributions and outliers, correlation vs causation, sampling bias, and how to read an A/B test result.

A typical case: "Variant B of a checkout page converted 150 of 2,400 visitors (6.25%) against 120 of 2,400 (5.0%) for the control. Should we launch B?" The relative lift is 25%, which sounds large. But a standard two-proportion z-test gives z ≈ 1.88 and a two-sided p-value of about 0.06, above the usual 0.05 threshold. A good answer says the result is promising but not conclusive at that threshold, and suggests running the test longer or agreeing the decision rule in advance.

Portfolio proof: in any project, report medians where data is skewed, state sample sizes, and add one line on what the data cannot tell you.

Python: increasingly expected, especially for cleaning

Python with pandas appears in many analyst job posts; for fresher roles, check whether each post lists it as required or preferred. It is most useful for cleaning messy files, combining sources and repeating an analysis every week without manual clicks.

Tests vary: some companies send a take-home dataset and ask for a notebook; others ask short questions such as "how would you find duplicate rows?" or "how do you merge two tables and keep unmatched rows?" Our pandas crash course covers the eight operations that come up most.

Portfolio proof: one notebook on GitHub that goes from raw file to finding, with markdown cells explaining each cleaning decision.

Business skills: the part tools cannot replace

This is where many technically strong candidates lose offers. Hiring managers want to see that you can start from a business problem, not from a dataset.

Three business skills matter most:

• Framing the question. Turn "sales are down" into "which cities and categories drove the 10% drop in orders between May and June, and was it fewer customers or smaller orders?"
• Choosing metrics. Know the basic metrics for the domain you apply to: conversion rate, average order value, retention, churn, customer acquisition cost, and so on.
• Breaking a metric down. For example, revenue = orders × average order value, and orders = visitors × conversion rate. When revenue falls, check each part in turn.

These are tested through case questions: "Our food-delivery orders in Bengaluru fell 8% last week. How would you investigate?" A strong answer checks data issues first (tracking, a holiday, an app outage), then splits the metric by segment, then forms hypotheses. Our data analytics interview questions guide has more of these.

Portfolio proof: every project README starts with the business question and ends with a recommendation.

Communication and working with AI tools

Every interview round tests communication, most directly in the project walkthrough. Practise a two-minute version of each project: the question, the data, what you did, the finding, the recommendation and one limitation.

AI assistants are now part of many analysts' daily work, for drafting SQL, debugging Python or summarising results. Interviewers increasingly care whether you can check AI output rather than just produce it. Be ready to explain every line of any query you show, and to spot when a generated answer uses the wrong join or the wrong date range.

Which skills should you learn first?

This order is ISS editorial guidance for someone starting from scratch. Adjust it to the job posts you are targeting.

Our data analyst roadmap lays this out month by month.

How do you prove these skills on a resume and portfolio?

Listing tools in a skills section is not proof. Hiring managers look for evidence that you applied a skill to a question and got a result. Aim for two or three complete projects that together cover the table at the top of this page:

• Project 1 (SQL + BI): an e-commerce or sales dataset, with SQL for a cohort or funnel analysis and a dashboard for a named audience.
• Project 2 (Python + statistics): a messy public dataset cleaned in pandas, with a finding that uses medians, segments and a clear limitation.
• Project 3 (Excel + business case): a workbook answering a specific business question, with a one-page recommendation.

Then write resume bullets that name the skill, the action and the result. Our data analyst resume guide for freshers shows bullet formulas and a full sample, and data analyst resume projects has more project ideas.

Quick self-check: are you interview-ready?

• Can you write a query with a CTE and a window function without looking it up?
• Can you build a PivotTable summary from a raw export in 15 minutes?
• Do you have one dashboard you can walk through in two minutes?
• Can you explain why a median can be better than a mean, with an example?
• Can you break a falling metric into its parts and list what you would check first?
• Can you explain every line of the code in your portfolio?

If you answered "no" to two or more, focus on those before applying widely.

Frequently Asked Questions

What are the top 5 skills for a data analyst?

SQL, Excel or Google Sheets, one BI tool such as Power BI or Tableau, basic statistics, and the business skill of framing a question and explaining a finding. Python is increasingly useful and worth adding once the first four are solid.

Do data analysts need coding skills?

Yes, at least SQL, which is a query language and is tested in most analyst interviews. Python is helpful for cleaning and repeatable analysis, but many analyst roles use it lightly or list it as a plus.

Is Python or SQL more important for a data analyst?

For most analyst roles, SQL comes first because company data lives in databases and SQL is the usual technical screen. Python adds value for cleaning messy files and automating repeated analysis.

How are data analyst skills tested in interviews?

Typically with a timed SQL test, an Excel or take-home task, case questions on metrics and statistics, and a walkthrough of a project you have built. Communication is judged in every round.

How can a fresher show data analyst skills without experience?

Build two or three complete projects on public data, each starting from a business question and ending with a recommendation. Publish the queries, notebook or dashboard and write resume bullets that describe the result.

Do I need a certificate to prove data analyst skills?

No. A certificate can help a resume pass an early screen, but interviews test what you can do. Projects you can explain in detail usually count for more.

Sources and methodology

This guide is ISS editorial guidance based on common analyst job-post requirements and interview formats. It contains no salary or job-count figures. All code was run in September 2026.

• The SQL example was run in SQLite 3 on the nine sample rows shown; the output is copied from that run. The query uses standard SQL that also runs in PostgreSQL and MySQL 8.
• The A/B test figures are a worked example on invented numbers, using a standard two-proportion z-test (pooled standard error), calculated in Python.
• Microsoft Support, SUMIFS function, checked September 2026.
• pandas, 10 minutes to pandas, checked September 2026.
• ISS: Data & Business Intelligence program page (week-by-week topics).

The learning order and time estimates are ISS estimates, not measured figures.

Next steps

Use the self-check above to find your two weakest skills and work on those first, with one project that proves each.

If you want a structured path through these skills with live sessions and feedback, review the Data & Business Intelligence curriculum. Applying is free, and you pay only after accepting an offer: apply here. You can also download the free Data Analyst Starter Kit on this page for a skills checklist and project briefs.