Skip to content
← All datasets
Dataset collection

Free Tableau Practice Datasets With Related Tables

Free CSV and Excel datasets for Tableau practice, with documented keys for relationships, dashboard ideas, and a checked total to confirm your data model.

Start with the ecommerce tables if you want to practice relationships: customers, orders, order lines, payments, and refunds, each at its own level of detail. Choose bike sharing for one hourly table you can chart in minutes, or workforce data for headcount and attrition views.

Ready-made practice

Files that load straight into Tableau

Start from one flat table for a first chart, or from a fact table and three dimensions when you want to practice a data model.

  • One flat sales table

    Every order line with customer and product fields already joined, for a first dashboard without a data model.

    Get the flat file →
  • Ecommerce star schema

    fact_sales with 237,537 order lines plus customer, product, and date dimensions as CSV files. Relate the three dimensions to the fact table.

    Get the star schema →
  • Learn fact and dimension tables

    Why measures stay correct when each table keeps its own grain.

    Read fact vs dimension tables →
Compare your options

Choose a dataset for the work you want to complete

DatasetConnect withData modelDashboard ideasMain challenge
Ecommerce Sales & ReturnsCSV files or Excel workbook6 tables to relateOrder activity, repeat customers, payments, and refundsPreventing duplicated revenue across order lines and refunds
SaaS Subscriptions & ChurnCSV files or Excel workbook5 tables to relateCurrent MRR, retention, billing, and payment outcomesDefining active status from one subscription period per account
Marketing Campaigns & ConversionsCSV files or Excel workbook4 tables to relateSpend, conversion rate, CAC, ROAS, and channel mixJoining daily spend to individual leads without fanout
Inventory & Supply ChainCSV files or Excel workbook7 tables to relateStock position, receipts, supplier lead time, and adjustmentsExplaining balances from opening stock and signed movements
Workforce & Employee AttritionCSV files or Excel workbook5 tables to relateHeadcount, recorded exits, tenure, role mix, and review coverageKeeping employee, employment-period, and review-event counts separate
AI Agent OperationsCSV files or Excel workbook8 tables to relateCost per session, completion rate, tool errors, and evaluation pass rateCombining call-level cost with session-level evaluations without double counting
Bike Sharing DemandCSV files or Excel workbookTwo grains of one table; use one at a timeHourly demand, seasonality, weather, and rider typeComparing hourly and daily grains correctly
US Flight DelaysCSV files or Excel workbook4 tables to relateOn-time rate, cancellations, delay causes, and route performanceAveraging delays when cancelled and diverted flights have none
Start here

Relationships for the ecommerce tables

Relate the six source tables on these keys, then check the model with one level of detail expression. Tableau may label customer_id as Customer Id; use the names your Data pane shows.

// Orders per customer, whatever dimensions are in the view
{ FIXED [Customer Id] : COUNTD([Order Id]) }
  • customers to orders on customer_id: one to many
  • orders to order_items on order_id: one to many
  • products to order_items on product_id: one to many
  • orders to payments on order_id: one to many
  • orders to refunds on order_id: one to many
  • refunds also carries payment_id. Relate it to orders or to payments, not both, because the data model cannot contain a loop.
  • Check for version 1.0.0: SUM(line_total) with cancelled orders excluded is 58,719,861.14.
  • Check for version 1.0.0: the highest value of the expression above is 16 orders for one customer.
Practical guidance

Start with a useful result

Relationships or joins

Use relationships for tables at different levels of detail, such as orders and order lines. A join merges the tables into one before anything is aggregated, so each order line would carry a copy of every payment on its order and the totals grow. Open a table and join inside it only when you want one flat table and the two sides match one to one.

A first dashboard worth publishing

Choose one question, for example which product categories drive sales and which lose the most to refunds. Build one view per measure, add a date filter that applies to all of them, and state the metric definitions in a caption. Tableau Public workbooks are visible to anyone, which suits practice data and a portfolio.

Where can I find free datasets for Tableau practice?

Every dataset on this page is free to download as CSV files and as an Excel workbook. Tableau connects to both. The ecommerce sales dataset is the best starting point for a data model: six related tables with documented keys and a checked total to compare your result with.

Do these datasets work with Tableau Public?

Yes. Tableau Public is free and connects to text files such as CSV and to Excel workbooks. Workbooks you publish there are public, which is fine for these practice datasets and a reason not to load private data.

Should I use relationships or joins in Tableau?

Use relationships when tables have different levels of detail. On the ecommerce data, the line total across orders that were not cancelled is 58,719,861.14. Joining order lines to payments before aggregating returns 62,015,901.22, because an order with more than one payment attempt repeats each of its lines.

Is a Tableau workbook included?

No .twb or .twbx file is included. You connect the CSV or Excel files and build the data source yourself, which is the part worth practicing. The dataset pages document every table, key, and relationship.

Understand the data behind the dashboard

Practice the joins, models, and definitions that turn source records into a report someone else can trust.

Explore data-modeling practice