Most people who stall on the way to a data analyst job do so because they learn tools in the wrong order, or try to learn all of them at once. This roadmap puts the four core skills in the order hiring managers actually test them, and shows what has changed in 2026 now that AI assistants sit inside Excel, Power BI and Tableau.
Facts in this article were checked against official vendor documentation in October 2026.
Why the order matters: Excel, SQL, BI tool, Python
A typical analyst job in India, whether at an IT services company, a bank, an e-commerce firm or a global capability centre, follows the same loop: pull data, clean it, summarise it and explain it to someone who has to make a decision. Each tool on this roadmap covers one part of that loop.
- Excel teaches you to think in rows, columns, lookups and pivots. Almost every business team still shares data in spreadsheets.
- SQL is how you get data out of company databases and warehouses. It is the most commonly tested technical skill in analyst interviews.
- A BI tool (Power BI or Tableau) turns your queries into dashboards that managers open every morning.
- Python with pandas handles what the others do badly: messy files, automation, larger datasets and basic statistics.
Learning them in this order means each new tool builds on the one before. Pivots in Excel prepare you for GROUP BY in SQL; SQL joins prepare you for data models in Power BI; and by the time you reach pandas, you already know what a “group by and aggregate” is supposed to return.
A six-month plan for working professionals
The plan below assumes about 8 to 10 hours a week alongside a job. Freshers studying full time can compress it to roughly half.
| Month | Focus | What “done” looks like |
|---|---|---|
| 1 | Excel for analysis | Clean a raw sales file, use XLOOKUP, build a PivotTable and a simple chart-based summary |
| 2 | SQL basics | SELECT, WHERE, GROUP BY, HAVING, all JOIN types, on a sample database |
| 3 | Intermediate SQL | CTEs, CASE WHEN, window functions (ROW_NUMBER, RANK, running totals), date handling |
| 4 | Power BI or Tableau | A three-page dashboard with a proper star-schema data model and a few measures |
| 5 | Python and pandas | Load, clean, merge and group data in a Jupyter notebook; plot results |
| 6 | Portfolio and job search | Three finished projects on GitHub, a one-page resume and interview practice |
Stage 1: Excel, now with AI inside it
Excel remains the first filter in many analyst interviews. Focus on the features you will use daily: tables, XLOOKUP, IF and IFS logic, text and date functions, PivotTables, conditional formatting and Power Query for repeatable cleaning.
Two newer features are worth knowing about, with caveats. Microsoft’s support pages say Python in Excel is available to Enterprise and Business Microsoft 365 users on Windows, Mac and the web, while Family and Personal users get it only as a preview. Copilot features in Excel, including the agent-style experience that builds and edits workbooks alongside you, need a Microsoft 365 Copilot licence or an eligible consumer plan. Do not build your learning plan around either. Many Indian employers have not rolled them out, and interviewers still want to see that you can do the work yourself.
Stage 2: SQL, the skill interviews test most
SQL is where most candidates are filtered out. Practise on a real database engine (PostgreSQL or MySQL are both free) rather than only in browser exercises. By the end of month three you should be able to write, without looking anything up:
- A query that joins three tables and filters on a date range
- A GROUP BY with HAVING that finds customers above a spending threshold
- A window function that ranks products within each category
- A CTE that calculates month-on-month growth
If your target company uses a cloud warehouse such as BigQuery, Snowflake or Microsoft Fabric, the SQL dialect differs slightly, but the concepts carry over completely.
Stage 3: Pick one BI tool and go deep
Choose based on the job listings in your city and industry, not on which tool looks nicer. Power BI is closely tied to the Microsoft 365 ecosystem that many Indian enterprises already pay for; Tableau is strong in organisations that use Salesforce. Learning both shallowly is less useful than learning one well.
Going deep means more than dragging charts onto a canvas. Learn data modelling (fact and dimension tables), relationships, calculated measures (DAX in Power BI, calculated fields and LOD expressions in Tableau), and publishing with row-level security.
Both vendors are now building AI agents and natural-language Q&A into their products. Microsoft’s September 2026 Power BI update made agentic authoring experiences generally available, and Tableau has been releasing conversational analytics features through 2026. These tools work best on top of a clean, well-documented semantic model, which is exactly the skill this stage builds. If you want a credential at this stage, we cover the PL-300 exam and Tableau certifications in separate guides.
Stage 4: Python and pandas
Python comes last because most entry-level analyst work can be done without it, but it is increasingly listed as “preferred” in job descriptions, and it is the bridge to data science roles later.
Note that pandas 3.0 was released on 21 January 2026, and its release notes describe meaningful changes: a dedicated string data type by default, copy-on-write behaviour (so chained assignment no longer modifies the original DataFrame), and a minimum of Python 3.11. If you follow an older tutorial, some code will behave differently. Make sure your course or book uses a current version.
Focus on: reading CSV and Excel files, filtering, groupby and aggregation, merge (the pandas equivalent of SQL joins), handling missing values, and basic charts with matplotlib or seaborn. Google’s own Data Analytics certificate now teaches Python rather than R, which tells you where the entry-level market has settled.
Stage 5: Portfolio, resume and interviews
Employers shortlist analysts on evidence. Three solid projects beat ten tutorial copies. A good set covers:
- An Excel or SQL project that answers a specific business question (for example, which product categories drive repeat purchases).
- A dashboard in Power BI or Tableau published with a short written summary of findings.
- A Python notebook that cleans a genuinely messy public dataset and explains each decision.
Write a short README for each project that states the question, the data source, the method and the result in plain English. That README is often what a recruiter actually reads.
Checklist before you start applying
- I can write joins, GROUP BY and window functions without notes
- I can build a PivotTable and use XLOOKUP and Power Query
- I have one published dashboard with a clean data model
- I can load, clean and summarise a CSV in pandas
- I have three projects on GitHub, each with a README
- I can explain one project end to end in under three minutes
Frequently asked questions
Do I need a degree in computer science to become a data analyst?
No. Many analysts in India come from commerce, economics, engineering, science and even arts backgrounds. What matters is demonstrable skill in SQL, Excel and a BI tool, plus the ability to explain findings clearly. Some employers do set degree requirements for campus hiring, so check individual job listings.
Should I learn Power BI or Tableau first?
Look at 20 to 30 job listings in your target city and industry and count which tool appears more often. Pick that one. The concepts of data modelling, measures and dashboard design transfer well, so switching later is not difficult.
Will AI tools make learning SQL and Excel unnecessary?
Not in the near term. Microsoft’s own Copilot in Power BI documentation warns that without preparation, Copilot can misinterpret data and return generic or inaccurate results, and Microsoft withdrew its experimental COPILOT worksheet function from Excel on 14 September 2026. Analysts are still needed to check AI output, and you can only check what you understand.
Official sources
- Microsoft Support: Python in Excel availability
- Microsoft Support: COPILOT function in Excel (retirement notice)
- Microsoft Learn: Copilot in Power BI overview
- pandas 3.0.0 release notes
- Power BI September 2026 Feature Summary
- Tableau new features (2026.2 release)
- Google Data Analytics Certificate
- PostgreSQL documentation: window functions tutorial

Leave a Reply