How to Become a Data Analyst: A Step-by-Step Roadmap

What does a data analyst do?

A data analyst takes a business question, finds the right data, cleans it, analyses it and presents the answer so that someone can make a decision. For example: Which of our 40 stores in Kerala and Karnataka lost sales this quarter, and why?

Most of the work is not advanced maths. It is getting data into shape, summarising it correctly and explaining what it means. That is why Excel users have a head start.

The roadmap

1. Get strong at Excel

   Excel is still the most common tool in analyst job descriptions, and it is where you learn to think in rows, columns and summaries.

   • Cell references, sorting, filtering and tables
   • SUMIFS, COUNTIFS, XLOOKUP, IF, and date and text functions
   • PivotTables and charts
   • Dynamic arrays: FILTER, SORT, UNIQUE
   2–4 weeks if you already use Excel

2. Learn to clean data with Power Query

   Real data is messy: extra spaces, merged headers, dates stored as text, data spread across many files. Power Query fixes these with repeatable steps that refresh with one click.

   • Import from CSV, folders and other workbooks
   • Split, trim, change type, unpivot
   • Merge and append queries
   2 weeks

3. Learn data modelling with Power Pivot

   The Data Model lets you connect several tables and analyse millions of rows inside Excel. This is the same engine used by Power BI, so everything you learn here carries over.

   • Import large files into the Data Model
   • Create relationships between tables
   • Write DAX measures such as total sales and year-on-year growth

   Start with Power Pivot →

   3 weeks

4. Learn basic statistics

   You need enough statistics to avoid wrong conclusions, not a degree.

   • Mean, median, mode and when each one misleads
   • Distribution, percentiles and outliers
   • Correlation versus causation
   • Sampling, and comparing two groups fairly
   2 weeks, alongside other steps

5. Learn SQL

   Company data usually lives in databases. SQL is how analysts pull exactly the rows they need.

   • SELECT, WHERE, ORDER BY
   • GROUP BY with SUM and COUNT
   • JOINs between tables
   • Subqueries and CTEs
   3–4 weeks

6. Build dashboards in Power BI

   Power BI uses Power Query and the Data Model you already know, and adds interactive reports you can publish and share.

   • Visuals, slicers and drill-down
   • Report design that answers one question per page
   • Publishing and scheduled refresh
   3–4 weeks

7. Optional: Python

   Python with the pandas library handles automation and analysis that is awkward in Excel. Learn it after the steps above, or when a job asks for it.

   Optional

8. Practise communication

   An analysis is only useful if people understand it. Practise writing a three-line summary of every analysis: what you found, why it matters and what to do next.

9. Build a portfolio

   Employers want proof. Build three or four projects with public data, for example from data.gov.in or Kaggle.

   • One Excel project: cleaned data, PivotTables and a dashboard
   • One SQL project: questions answered with queries
   • One Power BI report
   • Write a short post for each on LinkedIn explaining the question and the answer

10. Apply for the right roles

    Look beyond the title “Data Analyst”. Roles such as MIS Executive, Reporting Analyst, Business Analyst and Junior Data Analyst use the same skills and are good first steps.

Where AI tools fit in

AI assistants such as Copilot in Excel, ChatGPT and Claude can write formulas, explain DAX and draft SQL. They make a skilled analyst faster. They do not replace knowing whether the answer is right. Learn the fundamentals first, then use AI to speed up the work you already understand.

FAQ