Do data analysts need a portfolio website? A decision guide

What the reviewer is trying to learn

A hiring manager opening your portfolio link wants answers to three questions within a couple of minutes:

1. 1Can this person frame a business question?
2. 2Can they get a correct answer out of messy data?
3. 3Can they present it so I understand it without them in the room?

Any format that answers those three quickly is a good portfolio. Format matters less than whether the link opens on a phone, without a login, and shows the answer on the first screen.

The four options compared

A decision path

Work down the list and stop at the first "yes".

• Are your projects still weak or unfinished? Build projects first. A website around weak projects draws attention to the weakness.
• Are you applying mainly to analyst roles through portals and referrals? Use GitHub plus one public dashboard or PDF. This is the default for most people.
• Are you targeting roles where writing and stakeholder communication matter heavily, such as consulting-style analytics? Add two PDF case studies, linked from GitHub.
• Are you freelancing, or building a personal brand through writing? A simple website is justified. See first freelance clients for data analysts.

If you do build a site, keep it to four pages

Home         One paragraph about you, three project cards, contact link
Project 1    Case study: question, data, method, findings, limitations
Project 2    Same structure
About        Short background, tools, resume PDF link

Practical choices that keep upkeep low:

• Use a static site generator or a free hosting option that deploys from a GitHub repository, so the site updates when you push.
• If you buy a domain, set a renewal reminder. An expired domain on your resume looks careless.
• No blog section unless you will genuinely post. An empty blog with one post from last year signals the opposite of consistency.

The PDF case study format

This is the most underused option. Two to four pages per project:

Page 1  The question, who would ask it, and the decision it affects
        Data used (synthetic or public source) and its limitations
Page 2  Method in plain language, one key query or formula shown
Page 3  Findings: two or three charts, each with a one-line takeaway
Page 4  Recommendations, what you would check next, what the data cannot tell

Export from Word or Google Docs, check it opens properly on a phone, host it in the project repository, and link it in the README.

Data you must never publish

Whatever format you choose, never publish employer data, client data, customer lists, or anything with personal details, even "anonymised" by deleting a name column. Use synthetic data you generated, or public datasets whose licence permits reuse, and say which. Power BI's publish-to-web feature makes a report reachable by anyone with the link, so treat it as fully public.

Check it like a reviewer would

Send your portfolio link to someone who does not work in data. Ask them to open it on their phone and tell you, after two minutes, what one project found. If they cannot, simplify the first screen before you add more pages.

For the GitHub layer that every option depends on, use the GitHub profile README template. The Data Analyst programme produces dashboard and case-study projects on synthetic data that are safe to publish.