Recommended Free Tools
The best Power BI practice dataset is the one that gives you a useful business question and enough mess to solve it. A small financial workbook can teach measures and time-series analysis; a multi-table retail database can force you to build relationships; a transaction file can expose missing values, cancellations, and date problems that polished demos hide.
This list covers 13 datasets suited to Power BI practice, with an emphasis on finance, sales, profitability, budgeting, customer analysis, and the modeling skills that transfer well to personal-finance and business reporting.
Quick comparison
| Dataset | Best practice area | Format | Difficulty |
|---|---|---|---|
| Microsoft Financial Sample | Measures, filters, profit, time series | Excel or Power BI sample | Beginner |
| Retail Analysis Sample | Retail KPIs, drill-downs, inventory | PBIX or Excel | Beginner–intermediate |
| Sales and Marketing Sample | Funnels, campaigns, pipeline | PBIX or Excel | Intermediate |
| Human Resources Sample | Headcount, turnover, demographics | PBIX or Excel | Beginner–intermediate |
| Procurement Analysis Sample | Supplier spend and scorecards | PBIX or Excel | Intermediate |
| Customer Profitability Sample | Margins, variance, segmentation | PBIX or Excel | Intermediate |
| Maven Coffee Shop Sales | Transaction analysis and retail design | Excel | Beginner–intermediate |
| Maven CRM Sales Opportunities | Pipeline and win-rate analysis | CSV files | Intermediate |
| Maven NYC Taxi Trips | Large data, maps, performance | CSV and map files | Advanced |
| Maven Toy Store E-Commerce | Conversion and funnel measures | CSV files | Intermediate |
| Maven Global Electronics Retailer | Star schemas and currency conversion | CSV files | Advanced |
| UCI Online Retail II | Cleaning, RFM, cohorts | Excel | Intermediate–advanced |
| World Economic Indicators | Maps, trends, country comparisons | Public indicator data | Intermediate |
1. Microsoft Financial Sample
Best for: learning the basics without first spending hours repairing a file.
The Financial Sample contains sales and profit by country or region and market segment. It is a good first project for cards, slicers, clustered columns, maps, and line charts. Build measures for revenue, profit, profit margin, and year-over-year growth, then compare markets and countries.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minute#1 Best Overall
Microsoft provides the Financial Sample workbook. In Power BI Desktop, use Home > Learn with sample data or, on some start screens, Use sample data > Load sample data. In Navigator, choose Financials, then select Load.
2. Microsoft Retail Analysis Sample
Best for: retail dashboards, executive KPIs, drill-downs, and product analysis.
This sample combines retail sales, inventory, promotions, and customer behavior. Useful exercises include sales by store and product, inventory trends, promotional performance, and regional comparisons. Add a drill-through page for a store or product and practice conditional formatting on KPI tables.
The Microsoft samples page offers PBIX and Excel versions. The PBIX file includes the report, semantic model, queries, and visuals. The Excel version supplies data, not the finished dashboard.
3. Microsoft Sales and Marketing Sample
Best for: funnel analysis, campaign performance, sentiment, and opportunity tracking.
Use this dataset to calculate conversion rates, campaign return on investment, year-to-date performance, market share, and pipeline value. It is particularly useful if you want to practice separating activity measures from financial outcomes: leads and opportunities are not the same as revenue.
The Sales and Marketing sample is available as PBIX and Excel. Importing the Excel version will not recreate the finished sample report automatically.
4. Microsoft Human Resources Sample
Best for: workforce dashboards and demographic analysis.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
The sample includes headcount, turnover, diversity, workforce distribution, and employee-related KPIs. Try an attrition-rate measure, department and tenure bands, geographic analysis, and monthly hiring trends. For a finance-oriented exercise, connect headcount trends to an illustrative personnel-cost measure.
Use the version listed in Microsoft’s Power BI sample datasets. The figures are illustrative; they should not be treated as evidence about real workforce conditions.
5. Microsoft Procurement Analysis Sample
Best for: supplier scorecards, purchasing spend, quality metrics, and conditional formatting.
Practice spend by supplier and category, purchase trends, defect rates, supplier performance, and potential savings. A useful dashboard layout is a top-level spend and savings view followed by a supplier-detail page with risk and quality indicators.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →The PBIX and Excel versions are available through Microsoft’s sample datasets. Check the model before writing measures: supplier scorecards are only meaningful when the calculation’s grain is clear.
6. Microsoft Customer Profitability Sample
Best for: margin analysis, budget variance, customer segmentation, and financial storytelling.
This is one of the strongest choices for finance practice. Analyze revenue versus cost, profit by customer, budget variance, industry margin, business-unit performance, and year-over-year profitability. Build a waterfall showing how revenue turns into profit, then use segments and industries as drill-down dimensions.
Find it through Microsoft’s Customer Profitability sample. As with the other Microsoft examples, the Excel download contains source data while the PBIX version contains the complete report experience.
7. Maven Analytics Coffee Shop Sales
Best for: transaction-level analysis, time-of-day patterns, location comparisons, and product rankings.
The fictitious coffee shop operates at three New York City locations. Analyze sales by hour, weekday versus weekend performance, product contribution, and location-level results. This is a manageable way to practice heatmaps and transaction measures before moving to millions of rows.
The Maven Coffee Shop Sales page provides an Excel workbook with multiple tables and suggested analysis questions. Maven identifies the data as public-domain sample data, but check the current download page for access requirements.
8. Maven CRM Sales Opportunities
Best for: pipeline dashboards, win rates, salesperson comparisons, and quarter-over-quarter reporting.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →This fictitious computer-hardware company dataset includes accounts, products, sales teams, sales agents, and opportunities. Build a funnel from open opportunity to closed-won, calculate win rate, rank agents, compare products, and track pipeline value by quarter.
Download details are on Maven’s CRM Sales Opportunities page. The data is supplied as multiple CSV files and is listed as public domain, with the source identified as data.world.
9. Maven NYC Taxi Trips
Best for: large-file ingestion, geographic analysis, time intelligence, and model-performance testing.
The current listing covers green-taxi trips from 2017 through 2020, including pickup and drop-off times and locations, distance, fares, payment type, and passenger counts. Maven reports more than 28 million records and a download of about 403.7 MB.
Start with one year or one trip table. Load the full file only after your Power Query steps and model work. Suggested measures include trips by hour, average fare, average distance, and fare by route or borough. The Maven dataset page also links to map files and challenge guidance.
10. Maven Toy Store E-Commerce Database
Best for: website funnels, ecommerce relationships, and conversion measures.
This multi-table dataset includes sessions, page views, marketing channels, orders, and returns for a fictitious online retailer. Calculate session-to-order conversion, revenue per session, channel performance, customer acquisition measures, and product-launch results.
See the Maven Toy Store E-Commerce Database. The site currently lists an expanded version with more than one million records, so verify the selected download before quoting a row count. Some Maven pages display a sign-up or login requirement.
11. Maven Global Electronics Retailer
Best for: multi-table modeling, currency conversion, geographic reporting, and retail analytics.
The data covers transactions, products, customers, stores, and exchange rates for a fictitious global electronics retailer. Practice sales by country, product and store performance, customer segmentation, and reporting in a common currency.
Use a star-schema approach rather than joining the transaction table directly to every descriptive table. Confirm that product, customer, store, and currency keys are unique on their dimension sides. The Maven Global Electronics Retailer page provides the CSV files.
12. UCI Online Retail II
Best for: realistic cleaning, missing values, cancellations, customer metrics, RFM segmentation, and cohorts.
This real-world dataset contains 1,067,371 instances of UK online-retail transactions from December 1, 2009, through December 9, 2011. The download, online_retail_II.xlsx, is approximately 43.5 MB.
It is valuable because the problems are not hidden. Invoice numbers beginning with C indicate cancellations, CustomerID is missing for some transactions, and InvoiceDate includes a time component. In Power Query, create a date-only column before relating transactions to a calendar table. Decide explicitly whether cancelled invoices are excluded, shown separately, or netted against sales.
Build revenue, average order value, repeat-customer, monthly cohort, and recency-frequency-monetary measures. The UCI repository identifies the license as Creative Commons Attribution 4.0 International, so attribution is required.
13. World Economic Indicators
Best for: map visuals, country comparisons, time series, correlation analysis, and public-data storytelling.
Free tools Windows power users keep installed
One-click scans. No signup required.
Best Value
Indicators include measures such as population, GDP, human-development-related data, and life expectancy. Try a GDP-versus-population scatter plot, regional rankings, life-expectancy trends, or a map with a year slicer.
You can use Maven’s World Economic Indicators dataset or download indicators from World Bank Open Data. Values and definitions can change as the World Bank updates its files, so record the download date and the exact indicator definitions. Licensing and attribution requirements can differ between the Maven version and direct World Bank files.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to import these datasets into Power BI
Excel workbooks
- In Power BI Desktop, select Home > Excel workbook or Home > Get data > Excel workbook.
- Choose the workbook and select Open.
- In Navigator, select the relevant sheets or tables.
- Select Load, or choose Transform Data if columns need cleaning first.
For use in the Power BI service, Excel data should be stored in a table or model rather than being left as an unstructured range.
CSV files
- Select Home > Get data > Text/CSV.
- Choose the file.
- Check the delimiter and data-type preview.
- Select Load or Transform Data.
For NYC Taxi Trips, filter rows and remove unnecessary columns in Power Query before loading the model. This reduces memory use and makes early testing faster.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallRelationships in multi-table datasets
Power BI may suggest relationships, but do not accept them without checking the keys. Open Modeling > Manage relationships and use Autodetect only as a starting point. To define one yourself, select Manage relationships > New, choose the tables and columns, and check cardinality and cross-filter direction.
A normal one-to-many relationship requires unique values on the one side. Duplicate keys, mismatched data types, and ambiguous joins are common reasons autodetection fails. A clean star schema—fact table in the middle, dimensions around it—usually produces more reliable DAX than a web of many-to-many relationships.
A practical progression for learning
- Start with Financial Sample. Build cards, a date trend, a map, and three basic measures.
- Move to Customer Profitability or Retail Analysis. Add variance, segmentation, drill-through, and conditional formatting.
- Use UCI Online Retail II. Clean cancellations and missing customer IDs, then build cohorts or RFM segments.
- Choose a multi-table Maven dataset. Model dimensions and facts manually instead of relying on autodetection.
- Finish with NYC Taxi Trips. Test query folding where applicable, reduce the dataset, and compare report performance with different visual designs.
Mistakes that derail Power BI practice
- Assuming every sample includes a dashboard: PBIX files can include reports and models, but Excel versions generally provide only the underlying data.
- Trusting automatic relationships: inspect cardinality, uniqueness, data types, and filter direction.
- Joining dates with hidden times: a DateTime value such as
2020-01-01 14:35will not match a date-only calendar key. Remove the time component when the relationship requires dates only. - Loading the largest file immediately: begin with a filtered slice, especially on a low-memory computer.
- Using old 32-bit installation advice: Microsoft no longer supports the 32-bit version of Power BI Desktop; use the 64-bit version.
- Ignoring licensing: UCI Online Retail II requires CC BY 4.0 attribution. Public-domain labels do not automatically override the terms of an underlying source or every third-party download.
- Treating illustrative HR or profitability figures as real evidence: sample datasets are for modeling practice, not conclusions about actual companies or populations.
FAQ
Which Power BI dataset is best for a complete beginner?
Start with Microsoft’s Financial Sample. It is small enough to understand quickly and supports basic measures, filters, maps, profit analysis, and time-series visuals without requiring complex data cleaning.
Which dataset is best for practicing Power BI relationships?
Maven’s Global Electronics Retailer, Toy Store E-Commerce Database, and CRM Sales Opportunities are strong choices because each uses multiple CSV tables. Check key uniqueness and create relationships manually rather than assuming autodetect is correct.
What is the best dataset for finance and profitability dashboards?
Microsoft’s Customer Profitability Sample is the most direct choice. It supports revenue-versus-cost analysis, margins, customer segments, business units, industry comparisons, budget variance, and year-over-year profitability.
Can I use these datasets in a public Power BI portfolio?
Check the license and the download page for each dataset before publishing. UCI Online Retail II requires Creative Commons Attribution 4.0 attribution, while World Bank files can have source-specific requirements. Also describe synthetic Microsoft and Maven data accurately rather than presenting it as real company performance.
The Bottom Line
Use the Microsoft Financial Sample to learn the interface and basic DAX, Customer Profitability to practice finance storytelling, UCI Online Retail II to learn real-world cleaning, and Maven’s multi-table datasets to improve modeling. Leave NYC Taxi Trips until you are ready to manage large files and performance. The most useful project is not the one with the most rows; it is the one where you can explain the grain of the data, the relationship design, and why each measure produces the result shown.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
