Free Portfolio Projects for Data Analytics Beginners
Five free data analytics portfolio projects for beginners: public datasets, free tools, and free hosting for each, from first spreadsheet to first notebook.
You can build a complete data analytics portfolio for $0. Each of the five projects below pairs a public dataset you can download this week with a free tool and a free place to publish the finished work: a spreadsheet analysis of New York's 311 service requests, a Looker Studio dashboard on World Bank indicators, a SQL database built from the Stack Overflow Developer Survey, a Python notebook on real consumer complaints, and a Tableau Public dashboard on airline delays. You climb them in that order, because each rung adds exactly one skill an interviewer will test. None of it needs a paid course or a paid tool.
"Free" has to cover three layers before the claim is honest: the data, the tool, and the place the finished work lives. All three exist here. The datasets come from city, federal, and international portals that publish primary data. The tools are the free tiers professionals themselves use: Google Sheets caps a spreadsheet at 10 million cells (Google Workspace Updates), which is more room than most beginner projects ever need. And hosting is solved by share links, free notebooks, and free repository accounts. The same $0 logic carries a whole career change; the wider map of free courses, tools, and communities is in the free career change resources guide, and the exact limits of each free analytics tool are compared in the free data analytics tools review.
The ladder at a glancePermalink to “The ladder at a glance”
| # | Project | Free dataset | Free tool | Publishes to | Skill it proves |
|---|---|---|---|---|---|
| 1 | Clean messy city data | NYC 311 service requests | Google Sheets | Share link | Cleaning, pivot tables |
| 2 | Indicator dashboard | World Bank Open Data | Looker Studio | Public link | One-screen visualization |
| 3 | First real database | Stack Overflow Survey 2025 | SQLite + DB Browser | Free GitHub repo | SQL grouping and joins |
| 4 | First Python analysis | CFPB complaint database | Google Colab | Colab share link | pandas, a written verdict |
| 5 | BI build hiring managers open | BTS airline on-time data | Tableau Public | Public profile page | End-to-end BI delivery |
1. Clean and summarize messy city dataPermalink to “1. Clean and summarize messy city data”
New York publishes every 311 service request, from noise complaints to missed trash pickups, updated daily on NYC Open Data. The full file is too large for a spreadsheet, so you do what working analysts do: export a filtered slice, one borough and one quarter of complaint records.
Work it in Google Sheets. Deduplicate, split the date field, standardize the agency names, then build a pivot table that answers one question: which complaint types dominate this borough's records, and how long do they stay open? Write three sentences of findings under the table. The share link to that sheet, with the summary visible, is your first published artifact.
2. Turn one indicator into a decision dashboardPermalink to “2. Turn one indicator into a decision dashboard”
World Bank Open Data publishes development indicators for every country, free to download as CSV. Pick a single indicator with a story, internet penetration or health spending, and pull a twenty-year series for ten countries.
Build in Looker Studio, Google's free dashboard tool. The discipline here is subtraction: one screen, one trend view, one comparison view, one summary number, and a title that states the finding rather than the field names. When the public link opens on a phone and still makes the point, you have learned the thing dashboards exist to teach.
3. Build your first real databasePermalink to “3. Build your first real database”
The 2025 Stack Overflow Developer Survey received over 49,000 responses from 177 countries, and Stack Overflow publishes the full response set for free. Download the CSV, load it into SQLite with DB Browser for SQLite (both free, both standard practice), and you own a real database on day one.
Then ask it one question SQL is built for: do respondents who use AI tools at work report different job satisfaction than those who do not? Answering takes filtering, grouping, CASE logic, and honest handling of the "no answer" rows. Push the queries and a short data dictionary to a free GitHub repository. That repo is now the SQL section of your portfolio.
4. Run your first Python analysis in the browserPermalink to “4. Run your first Python analysis in the browser”
The Consumer Financial Protection Bureau publishes its Consumer Complaint Database: real, dated complaints about financial products, free to download. Open it in Google Colab, the free notebook environment that runs in a browser with no installation, and work it with pandas.
Keep the scope small: which products draw the most complaints, and what share of each gets a timely company response? One groupby, one percentage table, one chart, and a closing paragraph that says what a manager should do with the number. The Colab share link is the publish step, and the written verdict is what separates a project from an exercise.
5. Ship one BI build a hiring manager can openPermalink to “5. Ship one BI build a hiring manager can open”
The Bureau of Transportation Statistics publishes airline on-time statistics, with carrier and delay-cause detail, free for download. Load a year of it into Tableau Public, the free version of the BI tool hiring teams already use.
One screen: which carriers concentrate their late flights at specific airports, and how much of the delay is the carrier's own doing versus weather. This project joins flights to carrier lookups, designs for a single screen, and ends on a profile page. Your Tableau Public profile is a permanent, linkable portfolio that needs no explanation to open.
Where the ladder leadsPermalink to “Where the ladder leads”
Finished pieces from this ladder slot straight into the three-project portfolio that earns interviews, laid out in data analytics portfolio projects that get interviews. Career changers should then rebuild one rung on public data from their former industry, the move that makes a portfolio read like experience; the playbook is in portfolio projects for data analyst career changers. The destination pays like one worth the climb: data scientists earn a median of $120,230, with employment projected to grow 35 percent from 2025 to 2035 (BLS). Entry analyst pay sits lower, and the data analyst salary page tracks it level by level.
My own switch from sysadmin work into HealthTech analytics ran on the same budget. The project interviewers kept returning to lived in a free repository, built with free tools, and cost nothing to keep online. Nobody asked what the stack cost. They asked why I had excluded the bad rows, and that question is free to answer well.
How Traecta fits inPermalink to “How Traecta fits in”
Traecta — Your Personalized Career Roadmap sequences these five builds into a calendar instead of leaving them as a bookmark list: the spreadsheet week lands while the survey data is still fresh, each shipped project is logged as evidence against the analyst skill it proves, and the roadmap tells you when three finished pieces are enough to stop building and start applying.
Mistakes that keep free projects looking cheapPermalink to “Mistakes that keep free projects looking cheap”
- Publishing the tutorial version. A Titanic notebook proves you can follow instructions; a question you chose yourself proves you can work.
- Collecting ten half-finished dashboards. Three finished projects with written verdicts beat ten open tabs, every time they are read.
- Sharing files instead of links. A hiring manager will not request access, download a folder, or install anything. If it does not open in one click, it does not exist.
- Adding a sixth tool before finishing five projects. Tool breadth is what senior people show off; beginners are hired on depth.
What to take awayPermalink to “What to take away”
- The full path from zero to a data analytics portfolio costs $0: public data, free tools, link-based hosting.
- Climb in order: spreadsheet, dashboard, SQL, notebook, BI. Each rung adds one testable skill.
- Publish three finished projects with written verdicts, then start applying. Building more is procrastination with good branding.
If you want the sequencing and the stopping point decided for you, your personalized career roadmap from Traecta maps the five builds onto your weeks and tells you which three to show employers.
