Where can a Data Analyst find good, free public datasets for a portfolio project?

A data analyst can find free, high-quality public datasets for a portfolio project on platforms like Kaggle , which hosts hundreds of thousands of community-contributed files across every imaginable topic.

Here are the best places to find free public data:

• Kaggle : Large collection of user-uploaded datasets with built-in notebooks, community discussions, and clean CSV files.
• Google Dataset Search : A search engine built specifically to locate distributed datasets hidden across the web, academic repositories, and data hubs.
• Data.gov : The home of U.S. government open data, featuring thousands of topics including public health, crime, and climate.
• Maven Analytics Data Playground : Hand-picked, clean business and sports datasets specifically curated for data portfolio projects and monthly challenges.
• FiveThirtyEight GitHub : Real-world data behind the journalism site's engaging sports, politics, and culture articles, complete with clear README documentation.
• FRED (Federal Reserve Economic Data) : Massive repository of U.S. and international economic time-series data, perfect for forecasting and trend analysis.

If you'd like, let me know:

• What industry or topic you want to focus on (e.g., healthcare, finance, sports)
• Your preferred tool (SQL, Python, Tableau, Power BI)

I can recommend a specific dataset and suggest a project idea for your portfolio.

Absolutely. For a Data Analyst portfolio, I’d prioritize datasets that let you demonstrate more than just making charts—ideally ones where you can clean messy data, answer business questions, and build a compelling dashboard.

Best places to find free datasets

• data.gov — The U.S. government’s open-data portal. Excellent for transportation, healthcare, demographics, economics, crime, environment, and government operations. Great if you want a project that feels like real-world analyst work.
• kaggle.com — Huge variety of datasets covering sales, customer behavior, finance, sports, entertainment, marketing, and more. Particularly useful for portfolio projects because you can find datasets suited to specific business questions.
• archive.ics.uci.edu — A well-established academic repository with hundreds of datasets. It currently lists 689 datasets, including Bank Marketing, Wine Quality, and Heart Disease.
• ourworldindata.org — Fantastic for global economics, demographics, health, energy, climate, poverty, and population projects. Many datasets can be downloaded directly as CSV files, with metadata and source information included.
• datasetsearch.research.google.com — Useful when you have a particular topic in mind and want to search across many repositories.
• data.worldbank.org — Strong choice for economic and development analysis: GDP, employment, poverty, education, population, health, trade, etc.
• fred.stlouisfed.org — Excellent for economic and financial time-series projects. You can build analyses around inflation, unemployment, interest rates, housing, consumer spending, and similar topics.

What I'd choose for a portfolio

Instead of picking the most interesting-looking dataset, pick one that lets you tell a business story.

For example:

A particularly strong portfolio formula

Dataset → business question → SQL analysis → cleaned dataset → dashboard → recommendations

For example:

"What factors are driving declining customer retention, and which customer segments should the company prioritize?" Then demonstrate:

• SQL queries
• Data cleaning
• KPI calculations
• Customer segmentation
• Trend analysis
• An interactive Power BI/Tableau dashboard
• 3–5 concrete business recommendations

That will generally make a stronger portfolio piece than simply downloading a dataset and producing a collection of pretty charts.

If you're building your first Data Analyst portfolio, I can also give you 10 specific datasets + a project question + SQL questions + dashboard idea for each, ranging from beginner to advanced.

Absolutely. For a Data Analyst portfolio, I’d prioritize datasets that let you demonstrate more than just charts—ideally cleaning, SQL/Excel, exploratory analysis, visualization, and business recommendations.

Best free dataset sources

• kaggle.com — Probably the easiest place to start. Huge variety across retail, finance, marketing, sports, healthcare, and more. Look for datasets with enough rows to analyze but clear column definitions.
• data.gov — Excellent for real-world analyst projects. It currently catalogs hundreds of thousands of U.S. government datasets, covering transportation, health, economics, crime, education, climate, and more.
• ourworldindata.org — Great for compelling dashboards and trend analysis involving population, economics, energy, climate, health, and poverty.
• data.worldbank.org — Particularly good for international comparisons and economic/development analysis.
• archive.ics.uci.edu — A well-established academic repository with hundreds of datasets. Good for practicing structured analysis, though some datasets are more ML-oriented than business-oriented.
• datasetsearch.research.google.com — Useful when you have a specific topic in mind and want to discover datasets hosted by universities, governments, nonprofits, and other organizations.
• data.fivethirtyeight.com — Good for projects where you want to combine analysis with storytelling; datasets are associated with published data-journalism work.

What I'd choose for a portfolio

Rather than picking the "most interesting" dataset, pick one that resembles the job you want:

A particularly strong portfolio project might look like:

"Why are sales declining?"

Then use a public retail dataset to:

1. Clean and transform the data with SQL/Python/Excel.
2. Calculate KPIs such as revenue, profit, margin, and repeat-customer rate.
3. Identify trends by product, geography, customer segment, and time.
4. Build a Power BI/Tableau dashboard.
5. Finish with 3–5 concrete business recommendations.

That is generally more impressive than simply saying "I analyzed the Titanic dataset."

One important tip

Don't automatically choose the cleanest dataset. For a portfolio, a dataset that requires joining tables, handling missing values, creating calculated fields, and answering a realistic business question gives you much more to demonstrate. Recent portfolio guidance similarly recommends moving from clean beginner datasets toward messier government/real-world data as your skills improve.

If you're building your first Data Analyst portfolio, I'd start with Kaggle → Data.gov → Our World in Data, in that order.

Don't automatically choose the cleanest dataset. For a portfolio, a dataset that requires joining tables, handling missing values, creating calculated fields, and answering a realistic business question gives you much more to demonstrate. Recent portfolio guidance similarly recommends moving from clean beginner datasets toward messier government/real-world data as your skills improve.

A data analyst can find free, high-quality public datasets for a portfolio project on platforms like Kaggle, Google Dataset Search , and Data.gov.

Other Great Free Sources

• Data.world : A collaborative data community with thousands of open, user-contributed datasets tailored for social science and tech projects.
• GitHub Awesome Public Datasets : A massive curated list of high-quality topic-focused datasets categorized by subject like biology, economics, and machine learning.
• World Bank Open Data : Global development data, demographics, and international indicator metrics.

Would you like recommendations for datasets suited for a specific industry (like finance, healthcare, or e-commerce) or a particular tool (like SQL, Python, or Tableau)?

Absolutely. For a Data Analyst portfolio, I'd prioritize datasets that let you demonstrate more than just making charts—ideally data cleaning, SQL, exploratory analysis, visualization, and business-style recommendations.

Best free dataset sources

• data.gov — Excellent for real-world U.S. government data: demographics, healthcare, crime, transportation, housing, education, economics, and more. It currently catalogs 550,000+ datasets, so there's plenty to explore.
• kaggle.com — Probably the easiest place to start. It has hundreds of thousands of public datasets, plus notebooks showing how other analysts work with them.
• ourworldindata.org — Great for global topics such as population, climate, energy, poverty, health, and economic development. Particularly good for visualization projects.
• census.gov — Excellent for demographic, income, housing, employment, business, and geographic analysis.
• fred.stlouisfed.org — Great for economic and financial analysis: unemployment, inflation, interest rates, GDP, housing, consumer sentiment, etc.
• data.fivethirtyeight.com — Particularly useful because the datasets are connected to real analytical stories. That makes them good practice for learning how to turn data into a narrative.
• archive.ics.uci.edu — A classic source for well-known datasets. Better suited to statistical/ML-oriented analysis than business dashboards.
• data.worldbank.org — Excellent for international economic, demographic, education, and development comparisons.
• datasetsearch.research.google.com — Useful when you have a very specific topic in mind and want to search across many repositories.

What makes a dataset good for a portfolio?

Don't automatically choose the biggest dataset. A medium-sized, messy, realistic dataset can produce a much stronger portfolio project.

Look for data with:

• Multiple related tables → lets you demonstrate SQL joins
• Missing or inconsistent values → demonstrates data cleaning
• Dates → enables trend/time-series analysis
• Categories and numerical measures → enables meaningful segmentation
• Enough rows to make the analysis interesting
• A real-world question you can answer
• Some ambiguity that requires analytical judgment

For example, instead of "Analyze this Titanic dataset," a stronger project question would be:

"Which factors are most associated with customer churn, and what customer segments should a company prioritize for retention?" That gives you a reason to clean the data, explore relationships, segment customers, visualize findings, and make recommendations.

Portfolio-project ideas

My recommendation: If you're building your first serious analyst portfolio, start with Data.gov, Kaggle, Census, or FRED. They give you enough complexity to demonstrate actual analyst skills rather than simply producing a pretty visualization.

If you tell me whether you're targeting entry-level Data Analyst, BI Analyst, or Business Analyst roles, I can also suggest 5 specific datasets + a portfolio project question for each.

Finding the right dataset is the first hurdle for any standout portfolio project. Here are the top free public platforms where data analysts can find clean, interesting, and diverse datasets:

• Kaggle: The gold standard for beginners and pros alike. It hosts hundreds of thousands of public datasets across every niche imaginable, complete with user-written exploratory notebooks and discussions. Explore the Kaggle Datasets repository to filter by topic, file type, or usability score.
• Data.gov: The home of the U.S. Government's open data. Perfect if you want to tackle civic, economic, healthcare, or environmental questions. Check out Data.gov for massive repositories on public spending, climate trends, and safety metrics.
• Google Dataset Search: A powerful search engine built specifically to index datasets from across the web. You can use Google Dataset Search to find academic data, corporate disclosures, and niche repository links that aren't on mainstream platforms.
• GitHub (Awesome Public Datasets): A curated, community-driven repository listing high-quality data sources grouped by subject (finance, machine learning, sports, etc.). Browse the collection via the GitHub Awesome Public Datasets Repository.
• UCI Machine Learning Repository: A classic resource hosted by the University of California, Irvine, packed with clean, structured multivariate and time-series data. Visit the UCI Machine Learning Repository for traditional statistical and predictive modeling projects.
• World Bank Open Data: An incredible resource for global development data, including economics, education, poverty, and public health indicators across hundreds of countries. Access the metrics directly at World Bank Open Data.

If you'd like, let me know:

• What domain or industry you are targeting (finance, healthcare, sports, marketing, etc.)
• What tool you plan to use (SQL, Python, Tableau, Power BI)

I can recommend a specific dataset and a project idea to go with it.

Building a standout data portfolio requires clean, interesting, and diverse data. Here are the best free public dataset repositories categorized by domain to help you find your next project:

• Kaggle - The gold standard for data science and analytics, offering millions of user-contributed datasets across every imaginable industry with built-in Jupyter notebooks and community discussions. Browse through the Kaggle Datasets Hub.
• Google Dataset Search - A powerful search engine specifically designed to crawl and index thousands of repository sites, making it easy to find niche or academic data. Search for topics on Google Dataset Search.
• Data.gov - The official open data portal for the U.S. federal government, featuring massive collections of data on climate, health, education, and finance. Explore federal records at Data.gov.
• UCI Machine Learning Repository - A classic academic repository maintained by the University of California, Irvine, excellent for structured tabular data and classification/regression projects. Check out the archives on the UCI Machine Learning Repository.
• GitHub (Awesome Public Datasets) - A community-curated, massive markdown list of high-quality public data sources categorized by topic (sports, economics, biology, etc.). Access the list directly via the Awesome Public Datasets GitHub Repository.
• World Bank Open Data - A premier source for global development data, international economics, population demographics, and societal indicators. Find global metrics on the World Bank Open Data Portal.
• FiveThirtyEight - Famous for data-driven journalism, they publish the clean CSV files and code behind their viral articles on politics, sports, and culture. Browse the data on the FiveThirtyEight GitHub.

If you'd like, let me know:

• What industry or domain you want to focus on (finance, healthcare, sports, tech, etc.)
• What tools you plan to use (SQL, Python, Tableau, Power BI)

I can recommend a specific dataset and suggest a compelling business question to answer with it.

Absolutely. For a Data Analyst portfolio, I’d prioritize datasets that let you demonstrate more than just making charts—ideally you can show data cleaning, SQL, exploratory analysis, business questions, and a dashboard/story.

Great free dataset sources

My top picks for a portfolio

If you're trying to get hired as a Data Analyst, I'd choose one of these project types:

1. Retail / e-commerce

• Sales by product/category
• Customer retention
• Average order value
• Profit margins
• Seasonal trends
• Geographic performance

2. Marketing analytics

• Campaign conversion rates
• Customer segmentation
• Cost per acquisition
• ROI by channel
• Email campaign performance

3. HR analytics

• Employee turnover
• Salary vs. tenure
• Department performance
• Hiring trends
• Factors associated with attrition

4. Public data

• Housing prices
• Crime trends
• Transportation usage
• Population changes
• Income inequality

5. Global trends Our World in Data is particularly good here because you can combine variables—for example, GDP + life expectancy + education + population—and investigate relationships over decades. Its datasets provide downloadable CSVs and metadata, which is convenient for reproducible analysis.

One important tip

Don't pick a dataset just because it's large.

A 20,000-row dataset with a compelling business question is usually a better portfolio project than a 5-million-row dataset where you simply calculate averages.

A strong project might look like:

Business question: Why are customer cancellations increasing?
Tools: SQL + Python + Tableau
Process: Clean data → investigate trends → segment customers → identify drivers → recommend actions
Deliverable: Interactive dashboard + 3–5 key findings + business recommendations

That demonstrates the skills employers actually look for.

If you're building your first Data Analyst portfolio, I'd especially recommend Kaggle, Data.gov, UCI, and Our World in Data. UCI currently has hundreds of datasets and includes practical examples such as Bank Marketing, Census Income, Student Performance, and Real Estate Valuation.

Absolutely. For a Data Analyst portfolio, I’d prioritize datasets that let you demonstrate more than just charts—ideally something with messy data, multiple dimensions, and a question that could plausibly matter to a business or organization.

Best free sources

My top 3 for a portfolio

1. Data.gov — best for a "real analyst" project

Instead of downloading a perfectly clean dataset, find something that requires cleaning and combining. Data.gov contains data from many federal agencies as well as links to state/local datasets.

Example project:

"What factors are associated with traffic accidents in North Carolina?"

You could combine:

• Traffic crash data
• Population/demographic data
• Weather data
• Geographic data

Then demonstrate SQL → data cleaning → exploratory analysis → visualization → recommendations.

2. Kaggle — best for getting started quickly

Kaggle has a massive variety of public datasets and is particularly convenient when you want to practice a specific skill.

Search for things like:

• customer churn
• ecommerce sales
• marketing campaigns
• employee attrition
• retail transactions
• supply chain
• hotel bookings

I'd avoid the extremely overused beginner datasets like Titanic unless you're specifically using them to learn a new technique.

3. UCI — best for structured analytical practice

UCI is particularly good when you want something manageable and well documented. Current examples include Bank Marketing, Wine Quality, Adult Income, Student Performance, and Real Estate Valuation.

For an analyst portfolio, Bank Marketing is especially interesting because you can frame it as a business problem:

"Which customer characteristics are associated with successful marketing campaigns?"

That gives you opportunities to calculate conversion rates, segment customers, identify high-performing groups, and make recommendations.

One important tip

Don't choose a dataset by asking "What dataset looks interesting?"

Choose it by asking:

"What decision could someone make from this data?"

A strong portfolio project might follow this structure:

Business question → data collection → cleaning → SQL analysis → KPI calculation → visualization/dashboard → insights → recommendations

That's much more compelling to an employer than simply:

"I downloaded a CSV and made a Power BI dashboard."

If you're building your first Data Analyst portfolio, I'd recommend doing 2–3 projects with different types of data: one business dataset, one public/government dataset, and one time-series or geographic project.

Building a standout data portfolio requires clean, interesting, and realistic data. Here are the top platforms where you can find great, free public datasets, organized by category:

• Kaggle : The gold standard for beginners and pros alike. It features hundreds of thousands of community-uploaded datasets complete with user notebooks, data dictionaries, and built-in discussion boards. Explore the collection directly on Kaggle Datasets.
• Data.gov : The home of U.S. government open data. Excellent for finding civic, economic, environmental, and public health data to tackle policy or societal questions. Search through government records at Data.gov.
• Google Dataset Search : A powerful search engine specifically designed for datasets. It indexes data repositories across the web, making it easy to track down niche subjects by keyword. Try searching on Google Dataset Search.
• UCI Machine Learning Repository : A classic academic resource hosted by the University of California, Irvine. It's ideal if you want to lean into predictive modeling, statistics, or classic classification/regression tasks. Browse the archive at the UCI ML Repository.
• GitHub (Awesome Public Datasets) : A curated, community-driven markdown list of high-quality data sources broken down by category (finance, sports, biology, etc.). Check out the repository on GitHub Awesome Public Datasets.
• World Bank Open Data : Perfect for global macroeconomics, development indicators, population demographics, and health statistics across nearly every country. Access the data via World Bank Open Data.
• FiveThirtyEight : Known for publishing the raw data behind their data-driven journalism and political/sports analyses. The CSV files are clean, well-documented, and great for storytelling. Find them on FiveThirtyEight GitHub.

If you want, tell me:

• What industry or domain you want to target (e.g., finance, healthcare, sports)
• Which tools you plan to use (SQL, Python, Tableau, Power BI)

I can recommend a specific dataset and outline a project idea for it.