A data analyst portfolio is two or three small projects that show you can take a real question, work it out from messy data, and explain the answer. For a junior role it often matters more than certificates, because it shows the work itself.
What employers look for
- A clear business question, not just "I analysed some data".
- Evidence you cleaned the data, not only that you charted it.
- The right tool for the job: a formula, a query or a dashboard.
- A short, plain-English answer and what you'd recommend.
- Work that's easy to open in one click.
Three beginner project ideas
1. Retail: which stores are running out of stock?
Use a sales and stock dataset. Work out weeks of cover per store and product, flag the ones below two weeks, and recommend where to send new stock. Tools: spreadsheets, then a simple dashboard.
2. Healthcare: why do patients miss appointments?
Use a public appointments dataset. Work out the no-show rate by day of the week, age group and how far ahead the booking was made, and suggest one change. Tools: SQL and a chart or two.
3. Data cleaning: before and after
Take a deliberately messy dataset, such as customer records with duplicates, blank emails and mixed city spellings. List every problem, fix it, and show the row counts and totals before and after. Tools: SQL or spreadsheets.
How to write each project up
Keep each one to a single page or post, in this order:
- The question. One sentence at the top.
- The data. Where it came from, how many rows, what each column is.
- The cleaning. What was wrong, and what you did about it.
- The analysis. The key formulas, queries or dashboard, kept short.
- The answer. Two or three sentences in plain English, with one or two charts.
- The recommendation. "Based on this, I would…" This line turns a school exercise into analyst work.
Where to find data
- Government open-data portals, which publish health, transport, population and business data.
- Kaggle, which has thousands of free practice datasets.
- Your own life or work, with anything personal or private removed first.
Pick data from the industry you want to work in. A retail project for a retail job is worth more than a generic one.
Where to put your portfolio
- LinkedIn. A post or a Featured item for each project. It's where recruiters already look.
- GitHub. Good for SQL and Python files, with a short README for each project.
- A shared dashboard or document. Make sure the link opens without a sign-in.
Put the links near the top of your CV, next to your skills. Certificates you've earned go there too.
Mistakes to avoid
- Only the famous datasets. Employers have seen the Titanic and Iris datasets hundreds of times.
- Charts with no answer. Every project needs a conclusion and a recommendation.
- Skipping the cleaning. It's most of the real job, so show it.
- Too many projects. Two or three good ones beat ten half-finished ones.
- Private data. Never share real customer or patient details.
Questions people ask
How many projects do I need?
Two or three is enough for a junior role, as long as each one is finished and explained.
Do I need GitHub?
Not necessarily. It helps for SQL and Python work, but a clear LinkedIn post or shared document is fine, especially for spreadsheet and dashboard projects.
Can a portfolio replace experience?
For junior roles, it often counts as experience: it shows you can do the work. It won't replace years of experience for senior roles, but it gets you into interviews.
Next, read How to Become a Junior Data Analyst for the CV and job search.