6 Power BI Projects for Beginners (With Free Datasets You Can Use Today)

The best Power BI projects for beginners are small, finished dashboards built on real, slightly messy data: a sales report, an HR attrition report, a retail model with several tables, a personal finance tracker, an analysis of Indian open government data, and a before-and-after redesign of a bad report. Each one should show a different skill (cleaning data in Power Query, building a star schema, writing DAX measures, designing a page someone can read in ten seconds) so that together they prove you can do the job, not just follow a tutorial.

I'm Gulab Chand Tejwani. I run TechNet Consultancy in Anand and have taught Power BI to working professionals and students for years, and I hold the Microsoft PL-300 certification and the Microsoft Certified Trainer credential. The projects below are the ones I ask learners to build, in roughly this order.

What makes a good beginner Power BI project?

A good practice project has three things: data that needs some cleaning, a question a manager would actually ask, and a finished report you can explain in two minutes. Most beginners skip the first and the last. They load a perfectly clean CSV, drag twenty visuals onto a page, and stop.

Interviewers notice. When I review portfolios, the ones that stand out are the ones where the candidate can say "the date column came in as text in three different formats, so I fixed it in Power Query like this" or "I moved this calculation from a calculated column to a measure because the report was slow".

So for each project, keep a short note of the problems you hit and how you solved them. That note is worth as much as the dashboard.

1. Sales performance dashboard (Microsoft Financial Sample)

Dataset: Microsoft's free Financial Sample Excel workbook, which you can download from the Power BI documentation on Microsoft Learn.

The question: Which products, countries and segments drive profit, and how is that changing month by month?

Skills it proves:

• Importing from Excel and setting correct data types
• Building a separate date table and marking it as a date table
• Basic DAX measures: total sales, total profit, profit margin, and month-over-month change
• A clean one-page layout with cards, a trend line and a matrix

This is the "hello world" of Power BI, and that's fine. Keep it tight. One page, five or six visuals, and a title that states the main finding (something like "Profit fell in Q4 even though sales grew" tells the reader far more than "Sales Dashboard").

2. HR attrition report

Dataset: Search Kaggle for "HR analytics attrition". Several public datasets include employee age, department, salary band, tenure and whether the person left.

The question: Where is attrition highest, and what do the people who leave have in common?

Skills it proves:

• Grouping and binning (age bands, tenure bands)
• Measures that calculate a rate correctly (leavers divided by headcount, not a simple count)
• Slicers and cross-filtering that let an HR manager explore on their own
• Handling sensitive data thoughtfully: no names, only aggregates

Attrition rate is a nice DAX lesson because the obvious formula is often wrong. Getting the denominator right teaches you how filter context works far better than any definition.

3. Retail data model with multiple tables (AdventureWorks or Contoso)

Dataset: Microsoft's AdventureWorks or Contoso sample data. Both come as several related tables: sales, products, customers, stores and dates.

The question: How do sales compare across product categories, regions and customer groups, and which stores are underperforming?

Skills it proves:

• Designing a star schema: fact table in the middle, dimension tables around it
• Setting relationships and understanding one-to-many and filter direction
• Measures that work across tables, such as sales per customer or average order value
• Drill-down from category to subcategory to product

If you only do one project properly, make it this one. Data modelling is where most self-taught Power BI users are weakest, and it's the area that causes slow, wrong reports at work.

4. Personal finance tracker from your own bank statement

Dataset: Your own bank or UPI statement exported as CSV or Excel. Remove account numbers before you share anything.

The question: Where does my money go each month, and am I spending more than last year?

Skills it proves:

• Cleaning genuinely messy data: merged cells, header rows, inconsistent descriptions
• Categorising transactions with conditional columns or a mapping table
• Combining several monthly files from a folder, so next month's statement loads automatically
• Year-over-year comparison with time intelligence functions

This one is personal, which makes it memorable in interviews. It also teaches the "combine files from a folder" pattern, which you'll use constantly at work when someone emails you a new Excel file every week.

5. Indian open government data analysis (data.gov.in)

Dataset: The Open Government Data Platform India, data.gov.in, publishes thousands of datasets: rainfall, crop production, school enrolment, road accidents, health indicators and more, often by state and year.

The question: Pick one you care about. For example: how has crop production in Gujarat changed over the last ten years compared with other states?

Skills it proves:

• Working with public data that has gaps, footnotes and changing column names
• Unpivoting wide tables (years as columns) into a long format Power BI can use
• Map visuals and state-level comparisons
• Writing a short, honest commentary about what the data does and doesn't show

Recruiters see endless Superstore dashboards. An analysis of Indian public data shows initiative and gives you something real to talk about.

6. Redesign a bad report

Dataset: Any report from projects 1 to 5, or a cluttered dashboard you've seen at work (with the data anonymised).

The question: How can the same information be understood faster?

Skills it proves:

• Visual hierarchy: the most important number in the top left, details below
• Choosing the right chart (a bar chart beats a pie chart with twelve slices)
• Consistent colours, fewer fonts, clear titles
• Tooltips, bookmarks or drill-through instead of cramming everything onto one page

Show the before and after side by side. This is the project that convinces a hiring manager you can build reports their leadership will actually read.

How should you present Power BI projects in a portfolio?

Publish each project with three things: a screenshot of the main page, a short write-up (the question, the data, the main problems you solved and the key finding), and the .pbix file or a link to a published version. A free GitHub repository works well for the files and write-ups.

Power BI Desktop is free to download. Publishing to the web or sharing reports inside an organisation needs the Power BI service and, for most sharing scenarios, a paid licence, so check Microsoft's current licensing before you rely on a public link. Screenshots and the .pbix file are enough for most interviews.

How do these projects map to the PL-300 exam?

The Microsoft PL-300 (Power BI Data Analyst) exam measures four skill areas: preparing the data, modelling the data, visualising and analysing the data, and managing and securing Power BI. Projects 2, 4 and 5 cover data preparation. Project 3 covers modelling. Projects 1 and 6 cover visualisation. The fourth area (workspaces, row-level security, sharing) needs some extra practice in the Power BI service, which is worth adding once the basics are solid. Microsoft Learn publishes the current skills outline for PL-300; check it before you book the exam, because Microsoft updates it regularly.

Learn with guidance or on your own

You can do all six projects on your own with free tools. If you want structure, TechNet offers two routes:

• Power BI online course & toolkit: a self-paced course with lessons, labs and sample datasets, including PL-300 study guidance. Good if you're disciplined and want to study in your own time.
• Power BI training in Anand or live online: instructor-led batches where a trainer reviews your projects and helps when you get stuck. Good if you learn faster with feedback and a timetable.

If you're still deciding, read why DAX matters in data analysis and how to choose a Power BI course.

FAQ

How long does a beginner Power BI project take?

A focused beginner project takes roughly a weekend: a few hours to understand and clean the data, a few to build the model and measures, and a few to design and polish the report. The retail data model (project 3) usually takes longer because of the relationships.

Do I need to know SQL before starting Power BI projects?

No. Power Query handles most cleaning without SQL. SQL becomes useful when your data lives in a database, which is common at work, so it's a good second skill to add. See SQL training.

Can Power BI projects help me get a job without experience?

They help you get interviews and give you something concrete to discuss. They don't replace experience or guarantee a job, but a portfolio of three or four well-explained projects is one of the strongest things a fresher or career-switcher can show.

Is Power BI Desktop free?

Yes. Power BI Desktop is a free download from Microsoft for Windows. Sharing reports through the Power BI service generally needs a Pro or Premium Per User licence.

Which project should I build first?

Start with the sales dashboard (project 1) to learn the basics, then do the retail data model (project 3). Those two cover the skills most employers test first.