What does the data analyst roadmap look like at a glance?
The order matters because each stage feeds the next.
Excel teaches you to think in tables and metrics. SQL gets data out of databases. A BI tool presents it. Python handles jobs that are too big or too repetitive for the others.
| Stage | Focus | Time at ~10 hrs/week | Project to finish |
|---|---|---|---|
| 0. Foundations | Business metrics, spreadsheet basics | 1 week | A metric map for one business you know |
| 1. Excel | Cleaning, lookups, PivotTables, charts | 3–4 weeks | Sales performance report |
| 2. SQL | Filtering, GROUP BY, joins, CTEs, window functions | 5–6 weeks | Customer or orders analysis in SQL |
| 3. BI tool | Power BI or Tableau: data model, measures, dashboard design | 4 weeks | Interactive business dashboard |
| 4. Python | pandas for cleaning, analysis and simple charts | 5–6 weeks | Exploratory analysis notebook on messy data |
| 5. Capstone and job search | End-to-end project, resume, interviews | 4+ weeks | One case study using SQL, a dashboard and a written recommendation |
Time estimates are ISS editorial guidance for a beginner. People with an engineering or commerce background often move faster through stages 1 and 2. Statistics basics (averages, medians, percentages, distributions, simple A/B test logic) run alongside every stage rather than as a separate block.
Stage 0: Why start with business metrics?
An analyst's job is to help someone make a decision. Before any tool, learn what common metrics mean: revenue, orders, average order value, conversion rate, retention, churn and customer acquisition cost.
Project: pick a business you use, such as a food-delivery or shopping app. Write down its likely goal, five metrics that track it and one question you would want answered. This one-page "metric map" makes every later project sharper.
Stage 1: Which Excel skills do analysts need?
Excel (or Google Sheets) is still where many business teams receive and check numbers. Aim for:
- Cleaning: TRIM, text-to-columns, removing duplicates, fixing dates
- Lookups: XLOOKUP (or VLOOKUP and INDEX/MATCH)
- Conditional sums: SUMIFS, COUNTIFS, AVERAGEIFS
- PivotTables and PivotCharts
- Clean, labelled charts and a one-page summary
Project: a sales report from a public dataset showing revenue by region and month, the top and bottom products, and three written observations. Excel for the web and Google Sheets are both free with an account. Our advanced Excel course guide lists the skills analysts are tested on.
Stage 2: How much SQL do you need?
SQL is the skill most analyst screening tests check first, so give it the most time. Go beyond SELECT and simple joins to CTEs, window functions and date handling. You should be able to write a query like this without looking anything up:
SELECT category,
COUNT(*) AS orders,
SUM(amount) AS revenue,
ROUND(AVG(amount), 0) AS avg_order
FROM orders
WHERE status = 'delivered'
GROUP BY category
HAVING COUNT(*) >= 4
ORDER BY revenue DESC;On the practice shop dataset from our learn SQL in 30 days plan, this returns Electronics (4 orders, ₹21,000, average ₹5,250), Fashion (4, ₹6,600, ₹1,650) and Grocery (4, ₹2,700, ₹675). If you can explain why HAVING is used here instead of WHERE, you are ready for the next level.
Project: a customer analysis on a public e-commerce dataset: monthly revenue, repeat-purchase rate, top categories per city, and month-on-month growth using LAG. Free tools: SQLite with DB Browser for SQLite, PostgreSQL, or the BigQuery sandbox (no credit card needed).
Prefer to follow this roadmap with mentors and a cohort?
The ISS Data & Business Intelligence program follows a similar order: business metrics, Excel, SQL, then Python and data cleaning in the first six weeks, with dashboards and a capstone later in the 16 weeks. Compare it with the stages here, and download the free Data Analyst Starter Kit on this page for project briefs.
View Data & Business Intelligence curriculum →Stage 3: Power BI or Tableau, and how deep to go?
Pick one BI tool, based on job listings in your target city. Both are in demand; our Power BI vs Tableau comparison covers the differences. Learn:
- Connecting and shaping data (Power Query in Power BI)
- A simple star-schema data model: one fact table, a few dimension tables
- Measures (DAX in Power BI, calculated fields in Tableau)
- Dashboard layout: KPIs on top, trends in the middle, detail and filters below
| Tool | Free option (checked September 2026) | Paid plan, if you need to share inside a company |
|---|---|---|
| Power BI | Power BI Desktop is a free download | Power BI Pro: ₹1,165 per user a month, paid yearly, GST extra |
| Tableau | Tableau Public is free; published work is visible to anyone | See the vendor's pricing page |
You do not need a paid plan to learn or to build a portfolio. Project: an interactive dashboard answering one clear question, such as "Where are returns and delivery delays hurting revenue?" See Power BI projects for beginners for ideas.
Stage 4: When should you learn Python?
Learn Python after SQL and a BI tool, not before. Many entry-level listings treat it as "good to have", and it makes more sense once you know what analysis you want to automate.
- Python basics: variables, lists, dictionaries, loops, functions
- pandas: reading CSVs, filtering,
groupby, merging, handling missing values, dates - Simple charts with matplotlib or seaborn
- Working in Jupyter or Google Colab, which is free to use
Project: take a messy public dataset, clean it step by step in a notebook, and write up three findings with charts. Explain each cleaning decision in a markdown cell; reviewers care about your reasoning as much as the code.
Stage 5: What should the capstone and job search include?
Combine the stages into one case study: a business question, SQL to extract and shape the data, Python or Power Query to clean it, a dashboard to present it and a one-page recommendation. This single piece often does more for your applications than several small exercises.
Then move to applications. Our guide to data analyst jobs for freshers covers titles to search, where to apply and a 60-day plan, and data analyst resume projects shows how to present your work.
What should you skip early in the roadmap?
Most beginners lose months to topics that do not help them get a first analyst job. Leave these for later:
| Skip for now | Why | When to revisit |
|---|---|---|
| Machine learning and deep learning | Rarely asked of entry-level analysts; it is a data science skill | If you later move towards data science |
| A second BI tool | Employers want depth in one; switching is quick later | When a job requires it |
| Both Python and R | One scripting language is enough at the start | If your team uses R |
| Big data tools (Spark, Hadoop) | These are data engineering tools | If you move towards data engineering |
| Advanced statistics theory | Basics cover most entry work | Before experimentation-heavy roles |
| Collecting certificates | Projects show skill more clearly than a list of badges | Add one if a target employer asks for it |
AI tools are the exception to "skip": use them from stage 1 to explain errors and suggest formulas or queries. But always check the output against the data. An analyst who cannot tell when an AI-written query is wrong is not much use to a team.
What does a good weekly routine look like?
- Weekdays (1–1.5 hours): one lesson plus practice problems in the current stage.
- Weekend (3–4 hours): project work, then a short write-up of what you found.
- Every week: post one small finding or chart on LinkedIn or GitHub. It builds a visible record and forces you to explain your work clearly.
If you are comparing this path with data science, read data science vs data analytics. For pay at each level, see data analyst salaries in India.
Frequently Asked Questions
What is the right order to learn data analyst skills?
A practical order is business metrics and Excel, then SQL, then one BI tool such as Power BI or Tableau, then Python with pandas, finishing with a capstone project. Learn basic statistics alongside every stage.
How long does it take to become a data analyst?
At about 10 hours a week, the roadmap in this guide takes roughly 5 to 6 months to reach job-ready skills with a portfolio. The job search can take additional time. These are ISS estimates, and people with related backgrounds often move faster.
Should I learn Python or SQL first for data analysis?
SQL first. It is the skill most analyst screening tests check, and it is quicker to become useful in. Learn Python after SQL and a BI tool, once you know which analysis you want to automate.
Should I learn Power BI or Tableau?
Learn one. Check job listings in your target city and pick the tool they mention more often. Power BI Desktop and Tableau Public are both free for learning, and switching tools later is quicker than learning the first.
Do data analysts need machine learning?
Not for most entry-level analyst roles. Machine learning is mainly a data science skill. Focus on SQL, Excel, a BI tool, Python basics and statistics first, and revisit machine learning later if you move towards data science.
Can I follow this data analyst roadmap for free?
Yes. Excel for the web, Google Sheets, SQLite, PostgreSQL, the BigQuery sandbox, Power BI Desktop, Tableau Public and Google Colab all have free options, checked September 2026. Paid courses mainly add structure, feedback and mentoring.
Sources and methodology
Tool costs and free options were checked on official pages in September 2026.
- Microsoft, Power BI pricing (India) (Desktop free; Pro ₹1,165 per user a month paid yearly, GST extra), checked September 2026.
- Microsoft, Microsoft 365 for the web (free Excel for the web with a Microsoft account), checked September 2026.
- Tableau, Tableau Public (free; published work is public), checked September 2026.
- Google, Colab FAQ (free to use; paid tiers exist), checked September 2026.
- Google Cloud, BigQuery sandbox; DB Browser for SQLite; PostgreSQL License, checked September 2026.
- ISS: Data & Business Intelligence program page (weeks 1–6 syllabus).
The stage order, time per stage and "skip for now" list are ISS editorial guidance, not provider figures. The SQL example was run in SQLite 3.54 on the practice dataset from our 30-day SQL plan, and the output shown is from that run.
Next steps
Start with stage 0 this week: write a one-page metric map for a business you know. If you want to work through the roadmap in a live 16-week cohort with mentor feedback and a portfolio capstone, review the Data & Business Intelligence curriculum.
Applying is free, and you pay only after accepting an offer. Apply here.