Data Analyst Career Roadmap for Beginners

Start with a question and a small table you can inspect. A data analyst needs to produce a trustworthy answer, explain how it was calculated and recognise what the data cannot establish. This roadmap is written for beginners and freshers in India, including students from non-IT backgrounds. It is a sequence of practice, not a timetable. Move to the next stage when you pass the check at the end of the current one, not when a number of weeks has passed.

Stage 0: Prerequisites

You need a laptop, a spreadsheet tool and arithmetic you are comfortable with. Percentages, ratios, averages and the difference between a sum and a count are enough to begin. If any of these feel shaky, revise them first. Khan Academy's statistics and probability section is free and covers mean, median, spread and simple distributions without assuming a maths degree.

You do not need to know programming, and you do not need a paid course. You do need regular time. Two focused hours on most days beats a long weekend once a month.

You are ready to move on when you can explain, in your own words, why an average can be misleading when a few values are very large.

Stage 1: Foundations, learning to read a table

Practise spreadsheet sorting, filtering, formulas, dates and pivot tables. Identify what one row represents. In a fictional orders table, an order can contain several line items; counting rows may not count orders. This single idea, the grain of a table, prevents many wrong answers later.

Create a data dictionary for one small dataset. For each column write down the meaning, the unit, the allowed values and what a blank means. Then check for missing values, duplicates and inconsistent units. Your first milestone is a clean table with a documented transformation log: what you changed, why, and how many rows were affected.

A free resource for this stage is Microsoft's Excel help and training pages, which cover formulas, lookups and pivot tables with short exercises. Google Sheets works equally well if you prefer it.

You are ready to move on when you can take an unfamiliar table, state its grain, list its quality problems and fix them while keeping a record.

Stage 2: First project, SQL on a real question

Learn SQL in a useful order: selection and filtering, then aggregation with GROUP BY, then joins, then window functions. Check row counts and totals before and after each join. Use small synthetic tables where you can calculate the expected answer by hand, so you know when the query is wrong.

For your first project, build a fictional shop with customers, orders, order lines and refunds. Answer one question: what were monthly net sales, and how did refunds affect them? Explain the treatment of partially refunded orders and incomplete months. Write the queries, the results and a short note on what you are not sure about.

Kaggle Learn's Intro to SQL and Advanced SQL courses are free and run in the browser. SQLBolt is a free set of interactive lessons if you want more repetition. The PostgreSQL tutorial is useful once you install a database locally.

You are ready to move on when you can find a join that doubled a total, explain why it happened and repair it without adding DISTINCT as a reflex.

Stage 3: Depth, a report someone can use

Choose one audience and one decision. A store manager might need to know which products explain a sales change this month. Build a report with the relevant measures, comparison periods and filters. Resist adding every available chart. Every element should help the reader answer the question.

Reconcile dashboard totals with a query or workbook. Ask a friend to answer the intended question using only the report, without you explaining it. Revise labels or definitions where they struggle. Then write a one-page memo: observation, evidence, limitations and the next question. "Revenue rose while gross margin fell" is a finding; "the discount campaign caused the change" needs additional evidence.

Power BI Desktop is free to download, and Microsoft Learn has free Power BI learning paths. Add a little Python with Kaggle Learn's Pandas course if you want to clean larger files, but do not let tools replace the question.

You are ready to move on when your report, your query and your memo all agree on the same numbers, and you can explain every number's denominator.

Stage 4: Interview readiness

Rehearse the story of your project in three minutes: the question, the data, one quality problem, the calculation and the finding. Practise live SQL on paper and in a browser. Prepare to be asked about missing values, joins, chart choice and an unexpected result. Our companion guide on data analyst interview questions for freshers goes question by question.

Keep answers honest. If you have not used a tool, say so and connect the concept to something you have done.

You are ready when a mentor or senior can question your project for twenty minutes and you can defend every step or admit clearly where it is weak.

Coming from BCom, BBA, BSc or BA

A commerce or management background is a real advantage in understanding what revenue, margin, receivables and inventory mean. Many freshers from these streams underestimate this. What you usually need to add is comfort with data quality, SQL and one BI tool. Science and arts graduates often bring strong reasoning or writing skills; the writing matters because findings have to be communicated.

Do not describe yourself as "non-technical" in interviews. Describe what you have built. A BCom graduate with a reconciled sales report and documented SQL has analytical evidence that many CS graduates do not.

Realistic first roles in India

Data analyst, MIS executive, reporting analyst and analyst-trainee roles are accessible entry points, especially in operations, finance, e-commerce, retail and services companies. Many of these roles are spreadsheet and SQL heavy rather than Python heavy. Titles vary a lot between companies, so read the responsibilities rather than the title. Business intelligence and senior analyst roles usually come after one or two years of this kind of work. We do not quote salaries or hiring numbers here because they change by city, sector and year; check current listings yourself.

Mistakes beginners make

• Collecting certificates and tutorials instead of finishing one project end to end.
• Learning five tools shallowly rather than spreadsheets, SQL and one BI tool well.
• Trusting a dashboard total without reconciling it against the source.
• Treating a correlation in a chart as a cause.
• Hiding uncertainty. Interviewers respect "I checked this but could not confirm that".

What to put in your portfolio

One finished project is worth more than six half-done ones. Include the raw synthetic data, a data dictionary, the cleaning log, the SQL files with comments, the report or dashboard file, screenshots and a one-page memo. Add a short README that states the question and the answer in two sentences. Use only synthetic or openly licensed data and never upload anything from an employer or internship.

How GullyMentor approaches it

On the data analytics track (/career-tracks/data-analytics) a mentor sets a practice project like the one above, reviews your queries and report in detail, and then rehearses interviews with you. We sell no courses; the free resources named here are enough to learn from. We do not guarantee jobs, interviews or salaries.

Sources & further reading

• PostgreSQL SQL tutorial
• Khan Academy: statistics and probability (free)
• Microsoft Excel help and training
• Kaggle Learn: Intro to SQL (free)
• Kaggle Learn: Advanced SQL (free)
• SQLBolt interactive SQL lessons
• Kaggle Learn: Pandas (free)
• Microsoft Learn: Power BI training (free)

Examples and practice recommendations are educational illustrations. Tool documentation and programme inclusions should be checked for your intended use.