The best data analyst portfolio projects start with a business question, use real, messy data, and end with a clear recommendation. Two or three complete projects is enough: one SQL analysis, one interactive dashboard and one Python cleaning or automation project. Avoid overused datasets such as Titanic, and write up each project so a hiring manager can grasp it in two minutes.

Key takeaways

• Question → data → analysis → insight → recommendation. Every project, every time.
• Use fresh, messy data; cleaning it is part of the skill being shown.
• Lead every write-up with the finding, not the method.
• Link each project from your CV with a one-line outcome.

The structure every portfolio project needs

1. The question: “Which product categories drive repeat purchases?”
2. The data: the source, size, time period and its limitations.
3. Cleaning: what was wrong and how you fixed it.
4. Analysis: the SQL, Python or BI work, with your reasoning.
5. Findings: three bullet points, each with a number.
6. Recommendation: what a manager should do next.
7. Next steps: what you’d analyse with more time or data.

8 data analyst portfolio project ideas

#	Project	Skills shown	Data source ideas
1	E-commerce repeat-purchase analysis	SQL joins, cohorts, retention	Public e-commerce datasets on Kaggle
2	Sales performance dashboard	Power BI or Tableau, KPIs, drill-down	Sample retail datasets
3	Public transport or air quality trends in your city	Time series, cleaning, maps	Government open data portals
4	Customer churn drivers	Segmentation, comparison, statistics	Telecom or subscription datasets
5	Job market analysis for data roles	Text cleaning, Python, charts	Public job-posting datasets
6	A/B test evaluation	Statistics, experiment design	Published A/B test datasets
7	Personal finance or habit tracker	Excel/Python automation, storytelling	Your own anonymised data
8	Workplace reporting automation	Python, scheduling, documentation	Your team’s process, with permission and anonymised

Project 8 is often the most powerful: it shows real-world impact (“cut weekly reporting from 4 hours to 15 minutes”), and it is how many analysts make an internal move.

How to present each project

• Title as a finding: “Repeat customers drive 61% of revenue but only 18% of orders.” Use your own finding, not this example.
• A one-paragraph summary at the top.
• One key chart and a link to the interactive dashboard.
• Code on GitHub with a clear README.
• Limitations stated openly. Hiring managers trust analysts who know what their data can’t show.

Common portfolio mistakes

• Only tutorials, with no original question
• Beautiful dashboards that answer nothing
• No written findings, so the reviewer must interpret your chart
• Sensitive or proprietary data published without permission
• Projects abandoned halfway through

Using AI on portfolio projects

AI assistants can speed up code and drafting. Be transparent about using them, check every number they touch, and make sure you can explain every step in an interview, because you will be asked. See how to use generative AI at work.

The skills behind these projects are covered in SQL for data analysis, Power BI vs Tableau and the data analyst roadmap.

Build a capstone with a mentor

Tutopiya is planning a live Data Analytics with Python & SQL cohort that ends in a graded capstone project reviewed by a practising analyst, built to go straight into a portfolio. It isn’t running yet. Join the data analytics waitlist. There’s nothing to pay.

Sources

• Kaggle datasets
• Google BigQuery public datasets
• World Bank open data

Frequently asked questions

How many projects should a data analyst portfolio have? +

Two or three strong, complete projects are better than many small tutorials. Each should show a different skill: for example one SQL-heavy analysis, one dashboard and one Python cleaning or automation project.

Where can I find datasets for portfolio projects? +

Good sources include government open data portals, Kaggle, Google BigQuery public datasets, the World Bank and your own life or workplace (anonymised and with permission).

Where should I host my data analyst portfolio? +

GitHub for code and SQL, Tableau Public or Power BI's publish-to-web for dashboards (using non-sensitive data only), and a simple portfolio page or LinkedIn posts summarising each project's findings.

Should I use Titanic or Iris datasets? +

Avoid them as headline projects. They are so common that they don't differentiate you. Use them to learn, then build portfolio projects on fresher, messier data tied to a real question.

Founding cohorts · nothing charged

Want to learn this live, with a mentor?

Tutopiya is planning small, mentor-led online cohorts in English test prep, data & AI, business & finance and digital careers. Tell us which one you want. The programmes with the most interest open first, and the people on the waitlist hear about them first.

See the planned programmes