How to Build a Data Analyst Portfolio With No Experience

October 2, 2026 · 12 min read

By Mazen AhmedCo-founder & Growth Product Manager at SeeraReviewed by Ahmed Hussein

Quick answer: You build a data analyst portfolio with no experience from three self-made projects on public data, each with a different job: one that cleans a messy dataset, one that answers a specific question with SQL or Python, and one that presents the answer in a dashboard. Put the files where a stranger can open them, describe each project by what the data showed rather than by business results nobody measured, and link all three from one page. It takes about 30 days of evenings. If you already have projects and need the page layout, use the data analyst project page template; if you are not targeting analyst roles specifically, start with the general guide to building a portfolio with no experience.

Man writing notes in a lined notebook beside an open laptop at a small table by a window

Image credit: RDNE Stock project / Pexels

How this guide was put together: the three student portfolios were picked from the public gallery of data portfolios published on Seera, and each was confirmed live on October 2, 2026. Everything said about them comes from their public pages. The example figures in the findings section are illustrative; the quotes and the labour-market reference are linked in the sources.

Key Takeaways

An employer reads job history to learn four things about an analyst. Each of them can be shown without a job, as long as you know which piece of the portfolio is standing in for which.

What job history would have told them What shows it instead
You can work with data that arrives broken A cleaning project on a raw public dataset, with the decisions written down
You can turn a vague question into an answer A question project: one question, one query file or notebook, one stated finding
You can explain the answer to a non-analyst A dashboard project plus a three-sentence plain-language summary
Someone else has checked your work Public files, a cited data source, and coursework or a job simulation that was graded

Table of Contents

What "No Experience" Means to Someone Hiring a Junior Analyst

It means you have not been paid to do the work yet. It does not mean you have nothing to show, and people hiring for junior roles know the difference. The closest category the US Bureau of Labor Statistics tracks is data scientists, and its Occupational Outlook Handbook lists the work experience required in a related occupation as "None". The entry requirement for an entry-level analyst role is a degree and demonstrated skill. A portfolio is how you demonstrate it.

The standard career guidance says the same thing. Coursera's guide for new analysts puts it this way:

If you're just starting out and don't yet have work experience as a data analyst, include projects you've completed on your own or as part of your coursework.

So the question is not whether self-made projects count. They do. The question is which projects, how many, and how to describe them so that the reader trusts what they see. Those three decisions are the rest of this guide.

Three Projects, Three Different Jobs

Most beginner portfolios contain several projects that all prove the same thing: that the author can follow a tutorial to a chart. Three projects are enough if each one proves something the others do not.

Project What it proves Typical tool What you publish
1. The cleaning project You can make unusable data usable and explain your choices Excel, Power Query or Python (pandas) Before and after files, plus a short log of every decision
2. The question project You can answer one specific question and show the working SQL The query file and the finding in one sentence
3. The dashboard project You can hand the answer to someone who does not code Tableau or Power BI A public, interactive dashboard and a three-sentence summary

1. The cleaning project

Pick a dataset that is messy on arrival. Government open-data portals, city transit logs and public health files are good sources because nobody has tidied them for you. Tutorial datasets are poor sources for this project because the cleaning has already been done. CareerFoundry's Emily Stevens gives the reason in her portfolio guide: "Data cleaning, prep and transformation is a real part of any data job. Plus, it's hard to get noticed if you're using data everyone else has."

The deliverable is the decision log, not the clean file. Write down each choice in a line: which rows you dropped and why, how you handled missing values, which columns you renamed, what you could not fix. Five to ten lines is plenty. An interviewer will ask "why did you do it that way?" and the log is your answer, written before they asked.

2. The question project

Start with a question a manager could plausibly ask, written as a full sentence: "Which product categories lose the most customers after the first order?" Then answer only that. The discipline of one question is what makes this project look like work and not like exploration.

Use SQL here even if you are more comfortable in Python. SQL appears in most junior analyst job descriptions, and a readable query with a join, a common table expression and a window function tells a technical reviewer more in ten lines than a long notebook does. Publish the query file and state the finding in one sentence at the top of the project.

3. The dashboard project

Take the answer from project two, or a new dataset, and build the version a non-analyst would use. One page. Three or four charts. A filter or two. A title that states what the viewer is looking at. Publish it somewhere it can be opened and clicked without an account, and write a three-sentence summary a marketing manager could read: what the dashboard covers, the main thing it shows, and what you would look at next.

When you are ready to lay each of these out as a full case study, the project page template covers the structure section by section. If you want more project ideas with datasets once these three are done, eight data analyst portfolio projects that get interviews has full specs for each.

Write Findings, Not Business Impact

This is the rule most beginner portfolios break, and it is the one an interviewer notices first.

Portfolio advice written for working analysts says to lead every project with a business result: revenue recovered, churn reduced, hours saved. That advice is correct for people who have a business. A self-made project does not have one. Nobody acted on your analysis, so nothing was increased, reduced or saved. If the project page says it was, the first interview question will be "who measured that?", and there is no good answer.

What a self-made project does have is a finding: something the data showed, with a number and a scope, that you can reproduce on request. A finding is a weaker claim than an impact and a much stronger position, because you can defend every word of it.

The figures in those examples are illustrative, to show the shape of the sentence. Yours come from your data.

If you want to say what a business could do with the finding, say so as a recommendation and label it: "If this pattern held in a real store, I would test moving the promotion calendar forward by a month." That sentence shows commercial thinking without claiming a result. The same goes for datasets you generated yourself. Say the data is synthetic, in the first paragraph, and nobody will hold it against you.

When we reviewed seven established practitioners for our roundup of data science portfolio examples, none of them led with a business metric either. What they all led with was something a stranger could click and check. That is available to you on day one.

Three Real Portfolios Built Before a First Analyst Job

These three portfolios belong to students and recent graduates. Each solves the no-experience problem in a different way. Disclosure: all three were built on Seera, the product we make. They are included because they are live, public pages you can open and inspect, and you can see more in the gallery of data portfolios published on Seera.

One project per tool: Harsha Kumar Pasala

Harsha Kumar Pasala studied electrical and electronics engineering, not data. The portfolio has exactly three projects and they map to the three tools a junior analyst job description asks for: an Excel KPI dashboard on supermarket transactions, a SQL analysis of customer behaviour, and a Power BI report on mobile sales. Each project links straight to the file in a public GitHub repository: the workbook, the query write-up and the Power BI file. A reviewer who doubts the Excel skills can open the workbook.

What to copy: the structure. Three projects, one skill each, every one linked to the working file.

A cited dataset and a stated finding: Benedict Chidiebere

Benedict Chidiebere is finishing a statistics degree. The analysis project looks at sleep patterns across 62 mammal species and names its source, the Allison and Cicchetti dataset from 1976. The project description states what the data showed: larger animals sleep less, and the dataset's danger index was the strongest predictor of sleep. It links to the code on GitHub and points to a public write-up of the results. Nothing in it claims a business outcome, because there was no business. It is a finding, a source and the working.

What to copy: name the dataset, state the finding in a sentence, and write it up somewhere public.

Coursework and internships used as evidence: P. Tara Bhramaramba

P. Tara Bhramaramba is a current AI and machine learning student aiming at analyst roles. The portfolio leads with Tableau work: a set of dashboards on housing sale prices, property features and geographic patterns, and a dashboard concept for rainwater harvesting that came out of an internship programme. A data analyst internship and several training programmes fill the section where job history would normally sit. The bio says plainly that this is a student aspiring to a career in data analytics.

What to copy: say what stage you are at, and treat structured programmes as the evidence they are.

What Goes Where Job History Would Go

A portfolio page with an empty experience section looks unfinished. You do not have to leave it empty, and you do not have to inflate anything to fill it.

If you are switching careers

You have something a new graduate does not: you already know what questions a business asks. A nurse knows why readmission rates matter. A store supervisor knows what a stock-out costs. A teacher has tracked attendance against results. Build at least one of your three projects on public data from the industry you are leaving, and say why you chose it. Then look back through your old job for anything you counted, tracked or reported, and list it under that job as analysis. It was.

What you should not do is retitle the old job. "Operations Supervisor" with two lines about the weekly report you built is more credible than "Data Analyst" at a company that never employed one.

A 30-Day Plan From Nothing to a Live Portfolio

This assumes five to seven hours a week alongside study or a job. The page goes live at the end of week one with a single project on it, and you add to it from there. A portfolio with one finished project is more use to you than a plan for three.

Week What you do What exists at the end
1 Choose a messy public dataset. Clean it. Write the decision log. Create the portfolio page. Cleaning project published; page live with one project
2 Write one question. Answer it in SQL. State the finding in a sentence. Question project published with its query file
3 Build a one-page dashboard. Publish it. Write the three-sentence summary. Dashboard project live and clickable
4 Fill the experience section. Write the bio. Ask one non-analyst to read the page and tell you what you found. Finished portfolio, link added to resume and LinkedIn

The week-four test matters. If a friend outside the field cannot tell you what each project found after two minutes on the page, the summaries need rewriting before the link goes on an application. For the bio, how to write a portfolio bio has a structure that works for people at the start of a career.

Where to Publish It

Use three free places, each for one thing.

The third one is where people stall, because a GitHub profile feels like it should be enough. It is enough for a technical reviewer. It is not enough for the recruiter who sees your application first and will not open a repository to work out what you did.

The page does not need to be built by hand. Seera reads a resume, including a student one with only coursework and projects on it, and turns it into a portfolio page you can edit and publish; the free plan covers one portfolio. The student portfolio builder is the place to start, and the annotated data analyst portfolio example shows what a finished page contains.

Turn your resume into a data analyst portfolio — free →

Frequently Asked Questions

Can I build a data analyst portfolio with no experience?

Yes. A data analyst portfolio needs evidence, not an employer. Three self-made projects on public data are enough: one that cleans a messy dataset, one that answers a specific question with SQL or Python, and one that presents the answer in a dashboard. Publish the files where a stranger can open them, write a short plain-language summary of each, and link everything from one page. The US Bureau of Labor Statistics lists no related work experience as a requirement for the occupation, and hiring teams for junior roles expect to see coursework and self-directed projects.

How many projects should a beginner data analyst portfolio have?

Three is the right number to start with. One project looks like a tutorial you followed, and six looks like a list nobody will finish reading. Three lets each project prove something different: that you can clean data, that you can answer a question, and that you can communicate the answer to someone who does not code. Add a fourth only when it shows a skill or an industry the first three do not.

What can I use instead of work experience in a data analyst portfolio?

Use anything that produced a real output you can show: self-made projects on public datasets, coursework and capstone projects, job simulations, internships, certificates that ended in a graded case study, and analysis you did in a non-analyst job. Career switchers should lead with the last one. A spreadsheet you built to track stock levels in a retail job is analyst work, even though the job title was not analyst.

Should I put business results on a self-made project?

No. A self-made project has findings, not business impact, because no business acted on it. Write what the data showed and how you know, for example that weekend orders were 22 percent larger than weekday orders across 50,000 rows. Do not write that you increased revenue or improved efficiency by a percentage, because an interviewer will ask who measured it. Stating a finding you can defend is more convincing than claiming an impact you cannot.

Where should I host a data analyst portfolio for free?

Use three free places, each for one job. Put code, queries and workbook files in a public GitHub repository. Put interactive dashboards on Tableau Public or publish them from Power BI. Then put one portfolio page on top that introduces you, summarises each project in two or three sentences and links to the files. That page is the link you put on your resume and LinkedIn profile, so a recruiter never has to navigate a repository to understand what you did.

Useful sources and further reading

About the author

Mazen Ahmed — Co-founder & Growth Product Manager at Seera

Mazen is co-founder and Growth Product Manager at Seera. He works on how job seekers, students, and career changers grow their careers online, and writes about hiring, standing out, and building a professional presence from scratch.

Reviewed by Ahmed Hussein, Founder of Seera · Software Development Engineer at Amazon.

Editorial standards →