Data Analyst Roadmap 2026 for India: Excel, SQL, Power BI, Then Python

The most reliable data analyst roadmap for 2026 in India goes in this order: Excel, then SQL, then a BI tool such as Power BI, then Python. Each stage builds on the one before it, and each one maps to a task you will actually do at work, so you can start applying for junior roles after the first three and add Python while you job hunt. Plan on four to eight months of steady practice if you study part time.

I'm Gulab Chand Tejwani. I teach data analysis, SQL and Power BI at TechNet Consultancy in Vallabh Vidyanagar, Anand, and online. I hold the Microsoft Certified: Power BI Data Analyst Associate (PL-300) certification and I'm a Microsoft Certified Trainer. This is the order I recommend to almost everyone who asks me where to begin, and below I explain why, and what "done" looks like at each step.

Why does the order matter so much?

Because most people quit at the wrong tool. I meet a lot of students who started with Python because a video told them it was the "real" skill. Three weeks later they are stuck on installing libraries and they still cannot answer a simple business question like "which region sold the most last quarter?".

Analysis is about questions and answers. The tools are just ways to get from one to the other. Excel lets you see the data and the answer in the same window, which is the fastest way to build intuition. SQL teaches you to ask the same question of a database with millions of rows. Power BI teaches you to present the answer so a manager can act on it. Python comes last because it is the most flexible and the least forgiving.

If you flip the order you can still get there. It just takes longer and hurts more.

Stage 1: Excel (3 to 5 weeks)

Almost every Indian company, from a family business in Anand to a bank in Mumbai, still runs on spreadsheets. Your first analyst task is very likely to arrive as an .xlsx attachment.

What to learn:

• Clean data: Text to Columns, TRIM, removing duplicates, fixing dates stored as text
• Lookups: XLOOKUP (in Microsoft 365 and Excel 2021 onwards) and the older VLOOKUP and INDEX/MATCH, because you will meet old files
• Summaries: SUMIFS, COUNTIFS, AVERAGEIFS
• PivotTables and PivotCharts
• Power Query for repeatable cleaning steps
• Basic charts that a busy person can read in five seconds

You are done with this stage when you can take a messy export of a few thousand rows, clean it, and produce a one-page summary with a PivotTable and two charts without looking anything up. Don't rush this part. Excel habits, good and bad, follow you into every tool after it.

Stage 2: SQL (5 to 7 weeks)

SQL is the skill that separates "good with Excel" from "data analyst". Company data lives in databases such as MySQL, PostgreSQL, SQL Server or a cloud warehouse, and SQL is how you ask it questions. It is also the most common technical round in Indian analyst interviews in my experience, so it pays off twice.

What to learn, roughly in order:

1. SELECT, WHERE, ORDER BY, LIMIT
2. Aggregates with GROUP BY and HAVING
3. Joins: INNER, LEFT, and when a join quietly duplicates your rows
4. CASE expressions for bucketing
5. Subqueries and CTEs (WITH clauses)
6. Window functions: ROW_NUMBER, RANK, running totals with SUM() OVER
7. Date functions, because every business question has a time period in it

Install PostgreSQL or MySQL on your laptop (both are free) and load a sample database. You are done when you can take a plain-English question such as "top three products by revenue in each city last month" and write the query in under fifteen minutes. My SQL training covers this end to end, and the background on why SQL keeps showing up in job descriptions is in why SQL matters.

Stage 3: Power BI (4 to 6 weeks)

Once you can get the right numbers, you need to show them. In India, Power BI is the BI tool I see most in job listings for analyst roles, partly because so many companies already pay for Microsoft 365. Tableau and Looker Studio are fine too. The ideas transfer.

What to learn:

• Power Query: the same cleaning ideas as Excel, but repeatable and connected to sources
• Data modelling: fact and dimension tables, a star schema, relationships, a proper date table
• DAX measures: totals, ratios, time intelligence such as year-over-year
• Report design: one clear question per page, titles that state the finding
• Publishing to the Power BI service and setting a refresh

Microsoft's own PL-300 exam is a good map of what employers expect. Its skills outline groups the work into preparing data, modelling data, visualising and analysing data, and managing and securing it. You don't need the certificate to get a job, but working through the official study guide on Microsoft Learn is a solid checklist. If you're comparing courses, I wrote a separate guide on how to choose a Power BI course, and the Power BI training page lists what we cover.

You are done when you have two finished dashboards you can explain in two minutes each. My list of Power BI projects for beginners has ideas with free datasets.

Stage 4: Python (6 to 8 weeks, can overlap with job hunting)

Python with the pandas library is where you go when Excel is too slow, SQL is too awkward, or you need to automate a report that someone currently builds by hand every Monday. It is also the bridge to statistics and machine learning if you later move towards data science.

What to learn for analysis (not software engineering):

• Core Python: variables, lists, dictionaries, loops, functions
• pandas: reading CSV and Excel files, filtering, groupby, merge, handling missing values
• Plotting with Matplotlib or Seaborn
• Jupyter notebooks, and explaining your steps in markdown cells
• Basic statistics: mean versus median, distributions, correlation and why it is not causation

You are done when you can reproduce one of your Excel or SQL analyses in a notebook, end to end, and someone else can read it. Our Python training is built around this kind of work.

What about statistics, AI tools and cloud?

Statistics runs through every stage rather than sitting in its own box. Learn it as you need it: averages and percentages in Excel, distributions when you build charts, sampling and significance when you start Python.

AI assistants like Copilot in Excel and Power BI, or ChatGPT for writing a first draft of a query, are useful now and you should learn to use them. But they make confident mistakes. An analyst who cannot read the SQL that a tool wrote cannot check it, and checking is the job. Learn the fundamentals first, then let the tools speed you up.

Cloud warehouses (BigQuery, Snowflake, Azure Synapse, Microsoft Fabric) use SQL. If your SQL is solid, picking up a new warehouse takes days, not months. Don't start there.

How long does this take in total?

For a working professional or a final-year student doing about 8 to 10 hours a week, four to eight months is realistic for the first three stages plus a start on Python. Full-time learners can go faster. I've seen people stretch it to a year because they kept restarting from zero with a new course. Pick one path and finish it.

A rough weekly rhythm that works for many of my learners:

• Two short sessions learning new material
• One long session building something with a real dataset
• Thirty minutes writing down what you built and what went wrong

That last habit is what turns practice into interview answers.

What should a fresher put in a portfolio?

Three to four finished projects that each show a different stage of this roadmap: an Excel cleanup and summary, a set of SQL queries answering business questions, a Power BI dashboard, and a Python notebook. Use public data such as India's Open Government Data platform (data.gov.in) or Kaggle datasets. Put them on GitHub with a short README for each.

Finished and explained beats ambitious and half done, every time.

Where can I learn this in Anand or online?

At TechNet we run the full sequence as our data analysis training, either in the classroom at Vallabh Vidyanagar, Anand, or live online. Learners from Nadiad and nearby towns usually join online or come to the Anand classroom; we don't have a centre in Nadiad. If you're nearby, the training programs in Anand page has the details, and you can read more about local options in my earlier post on data analysis training in Anand and Nadiad.

Whatever course you choose, ask to see the syllabus, ask who teaches it, and ask what you will have built by the end. A roadmap is only useful if you walk it.

FAQ

Is Excel still worth learning for a data analyst in 2026?

Yes. Most business data in India still passes through Excel at some point, and interviewers often test lookups, PivotTables and basic cleaning. Excel is also the quickest way to build a feel for data before you move to SQL.

Should I learn Python or SQL first?

SQL first, for most analyst roles. It is shorter to learn, it is tested in most analyst interviews, and it teaches you to think in tables. Python makes more sense once you can already answer business questions with SQL and want to automate or go deeper.

Do I need a certification to get a data analyst job?

No certificate is required, but some help you structure your learning. Microsoft's PL-300 (Power BI Data Analyst Associate) is widely recognised for Power BI roles. Employers will still ask to see your projects and test your SQL.

Can a non-IT graduate become a data analyst?

Yes. Commerce, science and management graduates often make good analysts because they understand the business side. You need comfort with numbers, patience for messy data, and the willingness to practise SQL regularly.

How many projects do I need before applying?

Three or four finished, well-explained projects are enough to start applying. Quality matters more than count: each project should answer a clear question and show your cleaning, analysis and presentation steps.