CadetX
Portfolio & Projects

11 Data Analyst Portfolio Project Ideas That Actually Get You Hired (With Datasets)

Your portfolio is your experience when you don't have a job title yet. Here are 11 project ideas that show employers what they actually want to see — each with a real, free dataset to start from.

KM
Kavin Muthukumar
Founder, CadetX UK · Updated September 2026

When you're applying for data roles with no formal experience, your portfolio does the talking. It's the proof that you can actually do the work, not just that you sat through a course. But most beginner portfolios look the same — the same Titanic dataset, the same tutorial dashboard — and they blend into the pile.

The projects below are chosen to do the opposite: show a genuine range of skills, tell a clear story, and give a hiring manager a reason to keep reading. They're grouped by the skill each one demonstrates, and every idea comes with a real, free dataset you can use to build it today.

One rule before you start: employers care about the insight you draw, not the size of your dataset. A small, clearly-explained analysis that answers a real question beats a huge dataset with no story every time.

SQL & Data Querying

Show you can get and shape the data

1

Sales performance analysis with SQL

Query a retail or e-commerce dataset to answer real business questions: which products drive the most revenue, how sales trend by month, and which regions underperform. Write clean, commented queries and summarise the findings a manager would act on.

Dataset: Retail / e-commerce sales datasets on Kaggle (kaggle.com/datasets)
2

Public spending or crime analysis (UK open data)

Use UK government open data to explore a genuinely local question — how spending varies by region, or how reported crime shifts over time. UK-specific analysis stands out to UK employers and shows you can work with messy, real-world public data.

Dataset: UK Government open data portal (data.gov.uk)
3

Global indicators comparison

Pull economic or health indicators across countries and compare them over time — GDP vs life expectancy, or education spend vs outcomes. This shows you can join and interpret multiple dimensions, not just read one table.

Dataset: World Bank Open Data (data.worldbank.org)

Data Cleaning & Preparation

Prove you can handle messy, real data

4

Data cleaning case study

Take a deliberately messy dataset and document the entire cleaning process — handling missing values, fixing inconsistent formats, removing duplicates. This is unglamorous but it's what analysts do most, and showing it clearly signals real-world readiness.

Dataset: Any "raw"/uncleaned dataset on Kaggle or Data.gov
5

Survey data analysis

Survey data is messy by nature — free-text answers, inconsistent categories, missing responses. Clean it, categorise it, and pull out the themes. Great for showing judgement, not just technical steps.

Dataset: Our World in Data, or open survey datasets on Kaggle

Dashboards & Visualisation

Show you can communicate findings

6

Interactive business dashboard (Power BI or Tableau)

Build a dashboard a stakeholder could actually use — clear KPIs, filters, and a logical layout. The skill on display isn't "making charts," it's deciding what matters and presenting it simply.

Dataset: Tableau Public sample datasets, or Maven Analytics Data Playground
7

Trend / time-series visualisation

Visualise how something changes over time — energy prices, weather, streaming trends — and annotate what's driving the shifts. Employers love a chart that comes with an explanation, not just a line going up.

Dataset: Our World in Data (ourworldindata.org)
8

Geographic / map-based analysis

Map data by region — property prices, population, or public services — to reveal patterns a table would hide. Spatial analysis is visually striking in a portfolio and instantly demonstrates a step beyond bar charts.

Dataset: data.gov.uk regional datasets, or World Bank by country

End-to-End Analysis

The projects that separate you from the pile

9

Full end-to-end case study

Take one question from raw data all the way to recommendation: source it, clean it, analyse it, visualise it, and write a short conclusion. This single project demonstrates the whole workflow and is often the strongest piece in a portfolio.

Dataset: Any Kaggle dataset in a domain that genuinely interests you
10

A/B test or experiment analysis

Analyse the results of an experiment — did version A or B perform better, and is the difference meaningful? This shows a grasp of basic statistics and the kind of decision-support work businesses actually pay analysts for.

Dataset: A/B testing datasets on Kaggle
11

Analyse something you personally care about

The one project that's unmistakably yours. Your fitness data, a hobby, a local issue — a question you genuinely want answered. Personal-interest projects read as authentic, show initiative, and give you something real to talk about in interviews.

Dataset: Your own exported data, or Google Dataset Search for the topic

Dataset sources referenced are free, publicly available platforms widely used for data portfolio work, current as of 2026. Availability and contents change over time — check each source directly.

How to Present Them So They Actually Land

The projects only work if a hiring manager can understand them in under a minute. For each one:

The Shortcut: Real, Company-Sourced Projects

Self-directed projects are valuable — but they share one weakness: they're not real. You chose the dataset, you set the scope, and an employer knows it. The projects that stand out most are the ones built on real business problems, because they prove you can deliver in an actual work context.

That's the gap CadetX closes. Instead of guessing what to build, you work on company-sourced projects and walk away with a portfolio that reflects real professional standards.

100% Free

Virtual Work Experience Programme

Build your portfolio on real, company-sourced projects with global peers — instead of another tutorial dataset. Free, remote, three months, with a certificate at the end.

Apply Free →
Paid · Placement Support

Data & AI Launchpad — Zero to Hired

Want the skills and the portfolio and placement support? This structured programme takes you from the fundamentals all the way to getting hired.

Start Training →

Frequently Asked Questions

How many projects should a data analyst portfolio have?

Quality matters more than quantity. Three to five strong, well-documented projects that show a range of skills — SQL, cleaning, visualisation, and an end-to-end analysis — are more effective than a long list of shallow ones.

Where can I find free datasets for portfolio projects?

Good free sources include Kaggle, the UK government open data portal at data.gov.uk, Data.gov, the World Bank Open Data portal, Our World in Data, and Google Dataset Search. Choose a dataset that lets you tell a clear story rather than the largest one available.

What makes a data analyst portfolio project impressive to employers?

Employers care about the business insight you draw, not the size of the dataset. A strong project asks a clear question, documents the cleaning and analysis process, and communicates findings simply. Real, company-sourced project experience stands out even more.

Do portfolio projects matter more than certificates?

For getting hired, yes. A certificate proves you learned the skills; a portfolio proves you can apply them. The strongest candidates have both, but a portfolio is what turns a "no experience" application into a competitive one.

KM
Kavin Muthukumar Founder, CadetX UK

Kavin is the founder of CadetX, a career-launch platform helping students build real, employer-valued experience in data and AI roles.