A data analyst portfolio needs three good projects, not ten. What gets you a reply is the question each project answers, not the tool you used to answer it.
Most portfolios fail the same way. They use the same practice datasets everyone else uses, they show a chart with no conclusion, and they never say what the company should do next.
Here are the five steps to build a data analyst portfolio that stands out, with free tools and real Indian data.
The short answer:
- Build a data analyst portfolio in five steps.
- Pick a business question first.
- Find data nobody else is using.
- Build three projects: SQL, a dashboard, and a deeper analysis.
- Write each one to be skimmed in thirty seconds.
- Host it free on GitHub Pages or Tableau Public.
Let’s understand each step in detail.
Step 1. Pick a question, not a dataset
Start with something a company would pay to know. Most beginners do the opposite. They find a dataset, then hunt for something to say about it. That order shows in the finished work.
A dataset gives you rows. A question gives you a conclusion. Compare these two project titles:
| Weak title | Strong title |
|---|---|
| Analysis of Zomato data | Which city has the worst delivery times, and why |
| Sales dashboard | Where our discounting lost money last quarter |
| Netflix data exploration | What a streaming service should buy next in India |
The right-hand versions promise an answer. Write your title before you write any code. Make it a sentence a manager would care about. If you cannot, you do not have a project yet.
How many projects should a portfolio have?
Three finished projects beat ten half-built ones. Cover three different skills, one query project, one dashboard, one deeper analysis. Depth is what an interviewer can question you about.
Step 2. Find data nobody else is using
Use Indian government open data instead of the usual practice files. That one choice separates your work from most entry level portfolio examples online.
Almost every beginner portfolio uses the same few datasets. Titanic passengers, Netflix titles, a sample superstore. A hiring manager has seen them hundreds of times. They signal that you followed a tutorial.
Two free, official alternatives:
- NDAP, the National Data and Analytics Platform from NITI Aayog. It launched in May 2022 “for open public use”. NITI Aayog says that “all datasets on the platform can be downloaded and merged freely” (Press Information Bureau).
- data.gov.in, the Government of India open data platform, which publishes datasets from central and state departments.
Both give you something local and real. Rainfall, crop prices, road accidents, school enrolment, power use. Pick a subject you can talk about for ten minutes, and you will interview better.
Step 3. Build three projects, not ten
Cover three different skills with three projects, because that is what an interview actually tests.
Project 1: a SQL project. Load a few related tables. Answer five business questions with queries, and show each query beside its result. Our list of SQL queries for data analysts covers the patterns interviewers expect.
Project 2: a dashboard. Build one in Power BI or Tableau. Keep it to one screen with four or five charts that answer the title question. Our guides to Power BI projects and Tableau projects for beginners show the shape of a good one.
Project 3: a deeper analysis in Python or Excel. Clean messy data, find a pattern, explain what caused it. This is where you show thinking rather than tool skill.
That mix covers the three things on almost every job advert. Querying, visualizing and analyzing. The rest of the skills a data analyst needs build on those three.
Step 4. Write each project so it can be skimmed
Assume nobody reads your code. A hiring manager opens your project, looks for thirty seconds, and decides whether to keep reading.
Use the same five headings on every project page:
- The question: what you set out to find out, in one line
- The data: where it came from, and how big it was
- What you did: cleaning and method, in three sentences, no code
- What you found: the answer, with one chart
- What you would do about it: your recommendation
That last heading is the one almost nobody writes. It is also the one that gets replies. Anyone can produce a chart. Saying what a business should do about it is the actual job.
Keep one chart per finding. If a chart needs a paragraph to explain it, the chart is wrong, not the reader. Our note on why data visualisation matters covers how to pick one.
Should I use AI tools in my portfolio projects?
Yes, and say so. Using AI to speed up cleaning or code is normal practice now. Hiding it looks worse than using it. What you must own is the question, the method and the conclusion.
Step 5. Put it where recruiters can find it
Host your data analyst portfolio on a free public link you can paste into an application. Two options cover almost everyone. Both cost nothing.
GitHub Pages turns a GitHub project folder into a website. GitHub calls it “a static site hosting service”, and it is free on public folders with a GitHub Free account (GitHub Docs). Your portfolio website then sits at yourname.github.io.
Tableau Public hosts dashboards free. Tableau says you can save your work “to your personal Tableau Public profile to share with professional networks or potential employers” (Tableau). One warning. It is a public platform, so never upload anything confidential.
Then do three small things that take an hour in total:
- Put the link in your CV header, not at the bottom
- Put it in your LinkedIn profile, in the About section
- Write one short post per project explaining the finding
| Where you put it | What it does |
|---|---|
| GitHub Pages site | One link holding everything |
| Tableau Public profile | Live dashboards a recruiter can click |
| LinkedIn About section | Found by recruiters searching |
| CV header | Seen in the first six seconds |
Conclusion
A data analyst portfolio works when somebody opens it, understands what you asked, and sees what you decided. Tools change every year. That structure does not.
Do one thing this week. Pick a question you actually find interesting, find one dataset on NDAP or data.gov.in that touches it, and write the title before anything else.
Then finish it properly and write the five headings. One project you can defend beats five unfinished ones. It also gives you something specific to discuss in a data analyst interview.
One honest limit. A portfolio gets you the conversation, not the offer. Pair it with our data analyst career path guide, and with realistic numbers from our data analyst salary in India guide.
Frequently asked questions
1. What should a data analyst portfolio include?
Three finished projects covering SQL, a dashboard and a deeper analysis. Each needs a question, the data source, your method, the finding and a recommendation. A public link you can paste into applications matters as much as the projects themselves.
2. How do I build a portfolio with no work experience?
Use free public data and answer a question you care about. NITI Aayog’s NDAP platform launched in May 2022 for open public use. It states that all its datasets can be downloaded and merged freely. No job is needed to start.
3. What are good data analyst portfolio projects for beginners?
Start with one SQL project answering five business questions, one single-screen dashboard, and one cleaning and analysis project. Pick subjects from Indian open data such as rainfall, crop prices or road safety, which almost nobody else uses.
4. Where can I host a data analyst portfolio website for free?
GitHub Pages publishes a website straight from a public project folder, at no cost on a GitHub Free account. Tableau Public hosts dashboards free. Many people use both, with GitHub holding write-ups and Tableau holding the visuals.
5. Do I need a GitHub portfolio as a data analyst?
It helps, because it gives you one link and shows the work is real. GitHub Pages turns that same project folder into a website at yourname.github.io. You get a code record and a readable site from one place.
6. What are good sql projects for a data analyst portfolio?
Take two or three related tables and answer real business questions: which customers left, which product line fell, which region costs most to serve. Show the query beside the result so a reader can check your logic quickly.
7. How long does it take to build a portfolio?
Most people finish three solid projects in six to ten weeks working part-time. One project usually takes a weekend to build and a few more hours to write up properly, which is the part that takes longer than people expect.
8. Should my portfolio projects use Indian data?
It helps if you are applying in India. Local data gives you context an interviewer shares, and it avoids the practice datasets every other candidate used. data.gov.in and NDAP both publish free government data covering most sectors.
9. What is the difference between a data analyst portfolio and a resume?
A resume claims you can do the work. A portfolio shows it. The resume gets read first, so put your portfolio link in the header, then let the projects answer the questions a one-page resume cannot.
10. Can a portfolio get me a data analyst job without a degree?
It can get you the conversation. Many Indian firms still screen on degrees and certificates. Treat the portfolio as the thing that proves you can do the work, once you are already in the room.
