10 data analyst portfolio project ideas that actually get you hired
Not random Kaggle notebooks. Ten data analyst portfolio projects that mirror the work analysts are paid to do, what each one proves to a recruiter, and how to make them interview-ready.
The problem with most data analyst portfolios is not too few projects. It is the wrong ones: a Titanic notebook, an iris classifier, three tutorials with the numbers changed. None of them look like the job, so none of them earn an interview.
A portfolio project earns its place when a recruiter can look at it and think "this person has done something close to what we would pay them for". Below are ten that clear that bar, grouped by what they prove. Pick three or four, finish them properly, and write up each one.
What a hireable project actually shows
Before the list, the test every project should pass. A strong data analyst portfolio project does four things: it answers a real business question, it uses a realistic (ideally messy) dataset, it shows your cleaning and reasoning rather than hiding it, and it ends with a recommendation a stakeholder could act on. If a project cannot survive the question "so what would you tell the business?", it is an exercise, not a portfolio piece.
Foundational projects: prove you can handle data
SQL business report
Take a real transactional dataset and answer a business question end to end in SQL: revenue by segment, the cause of a drop, top and bottom performers. Proves the one skill every analyst job requires.
Messy data cleaning case study
Take a deliberately messy dataset (duplicates, nulls, inconsistent labels), clean it, and document every decision. Proves you can be trusted with real data, which is never tidy.
Exploratory analysis with a verdict
Explore a dataset, but end with a clear finding and recommendation, not just charts. Proves you analyse toward a decision instead of describing numbers.
Spreadsheet to insight
Turn a raw spreadsheet into a clean model and a one-page summary. Proves you can deliver in the tool most businesses still run on, which matters more than juniors think.
Dashboard projects: prove you can communicate
These are the projects hiring managers click first, because the output looks like the deliverable they need.
Interactive BI dashboard
Build a dashboard in Power BI or Looker Studio that answers one decision (where is marketing spend wasted, which region is underperforming). Proves you can turn data into something a non-analyst reads in 30 seconds.
Web analytics funnel
Use the public GA4 sample in BigQuery to build a conversion funnel and find where users drop off. Proves you can work with event data and the cloud warehouse stack employers actually use.
Cohort and retention analysis
Group users by signup cohort and track retention over time. Proves you understand the metric every subscription and product business cares about most.
Executive KPI report
Build a single-page KPI view for a fictional leadership team, with clear definitions for each metric. Proves you can scope what matters and define metrics precisely, a senior skill.
Capstone projects: prove you can do the whole job
End-to-end analytics project
Raw data in, cleaning, analysis, dashboard out, written recommendation. The one project that mirrors a real analyst's week from start to finish. This belongs at the top of your portfolio.
A/B test or campaign analysis
Analyse an experiment or marketing campaign and state, with evidence, whether it worked and what to do next. Proves you can support a real business decision with data, which is the whole point of the role.
See what a finished portfolio looks like
Project ideas are easy. Knowing when one is "done" and recruiter-ready is the hard part. The clearest way to calibrate is to look at finished examples. These are real D8A portfolios built from exactly the projects above:
A full data analyst portfolio: SQL foundations, Python analysis, a BigQuery and GA4 project, and a Power BI dashboard, each one finished and documented.
“World's best boss of business intelligence. I make SQL look easy and hard truths sound like a 'that's what she said.'”
Open the portfolio →A business analyst portfolio built around requirements, KPI dashboards, and a business case, for the analyst-adjacent route.
“Business Analyst and professional stakeholder whisperer. I have a lot of feelings about the requirements doc.”
Open the portfolio →How to make any of these interview-ready
- 1
Lead with the question, not the tool
Write the business question at the top of every project README. Recruiters care what you answered, not which library you imported.
Questions to ask- What business decision does this support?
- Who would act on the answer?
- What changes depending on the result?
- 2
Show the mess, not just the result
Your cleaning and judgement calls are the most hireable part. Document them instead of hiding them behind a clean final chart.
Questions to ask- What was wrong with the raw data?
- What did you decide to do about it, and why?
- What did you choose to leave out?
- 3
End with a recommendation
Close every project with a recommendation. An analysis with no "therefore" reads as an exercise. One with a clear call reads as work.
Questions to ask- So what should the business do?
- What is the one sentence a manager would repeat?
- What would you check next?
- 4
Make it public and easy to open
A portfolio piece a recruiter cannot open does not exist. Host it, write the README, and make the first 30 seconds count.
Questions to ask- Is it on a public URL or GitHub?
- Can someone understand it without you in the room?
- Does it load in seconds?
A certificate says you watched a course. A portfolio says you can do the work, and recruiters can tell the difference. The ten ideas above are the raw material. Finishing three of them, documented and public, is what turns into interviews.
The fastest way to build them without stalling is a guided brief for each, which is exactly what each D8A path gives you: these projects, in order, auto-validated, and published to a portfolio as you go.
Frequently asked questions
- What projects should a data analyst put in a portfolio?
- Projects that mirror real analyst work: a SQL business report, a cleaned and analysed messy dataset, an interactive BI dashboard, a marketing or product funnel analysis, and an end-to-end capstone. Three to five finished, documented projects on one theme beat a dozen half-built notebooks.
- How many projects do I need in a data analyst portfolio?
- Three to five strong projects is enough. Recruiters do not count projects, they judge the best one or two and check you can explain your choices. One polished, well-documented project is worth more than five rushed ones with no write-up.
- What makes a data analyst portfolio project stand out?
- It answers a real business question, uses a realistic dataset, shows your cleaning and reasoning, and ends with a clear recommendation. A short write-up explaining the decision it supports matters more than fancy charts. Make every project something you can defend in an interview.
- Where do I find datasets for portfolio projects?
- Public sources like Kaggle, data.gov, the Google Analytics sample in BigQuery, and company open data work well. The dataset matters less than the question you ask of it. A simple dataset with a sharp business question beats a huge dataset with no clear point.
