How to Use Shared Dataset in Power BI Desktop

Doing the same work over and over is one of the most frustrating parts of data analysis. If you've ever found yourself rebuilding the same data model or recreating the same DAX measures for different Power BI reports, you know the pain. This article will show you how to solve that problem for good by using shared datasets in Power BI, allowing you to build your data model once and reuse it across countless reports.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

What is a Power BI Shared Dataset, and Why Should I Use One?

In a standard Power BI workflow, you connect to data sources, clean and transform them in Power Query, build a data model with relationships, write DAX measures, and design visuals - all inside a single PBIX file. This is great for one-off reports, but what happens when you need to create five different reports, all based on the same core sales data?

Without a shared dataset, you’d have to copy that file five times or, even worse, rebuild the model from scratch for each report. This is where the magic of shared datasets comes in.

A shared dataset is a Power BI data model that has been published to the Power BI Service and made available for other users to connect to. It acts as a single, centralized, and reusable data layer. You do the hard work of cleaning, modeling, and creating measures just once. Then, you and your colleagues can connect to this "master" dataset to create new reports that only contain the visuals, saving a massive amount of time and effort.

Here’s why this approach is a game-changer:

• A Single Source of Truth: Everyone uses the same verified data, relationships, and calculations. This puts an end to arguments over whose sales numbers are correct, because every report is built on the exact same foundation.
• Huge Time Savings: Why spend hours building the same data model again and again? Build it once, perfect it, share it, and then focus your energy on creating insightful visualizations.
• Unmatched Consistency: When all reports stem from one dataset, they automatically have consistent naming conventions, formatting, and DAX formulas. A "Total Revenue" measure will mean the same thing everywhere.
• Effortless Maintenance: Need to add a new sales metric or fix a bug in a calculation? Update the central dataset once, and the changes automatically flow through to every single report connected to it. No more need to update a dozen different files.
• Improved Data Governance: You can control who has access to sensitive data at the dataset level. Report creators can build visualizations without needing direct access to the underlying SQL database or raw data source.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

The Two-Part System: "Golden Datasets" and "Thin Reports"

The shared dataset methodology revolves around two key concepts. Understanding them will make the whole process click.

1. The Golden Dataset (or Master Dataset)

This is the cornerstone of the entire system. The "Golden Dataset" is the PBIX file where you build your perfect, authoritative data model. This file contains everything except the visuals for a final report. Its only job is to be the pure, well-structured source of truth.

Inside this file, you will:

• Connect to your data sources (SQL, Excel, SharePoint, etc.).
• Perform all your data cleaning in the Power Query Editor.
• Establish all table relationships in the Model view.
• Write all your core DAX measures (e.g., Total Sales, YTD Revenue, Customer Count).

Once this model is complete and verified, you publish it to a workspace in the Power BI Service. It then becomes the "golden" source that others will use.

2. Thin Reports

A "Thin Report" is what you create when you connect to a Golden Dataset. It’s called "thin" because it is a lightweight PBIX file that contains only the report elements - visuals, themes, bookmarks, and report-level measures. It does not contain its own data model, tables, or relationships.

Instead of connecting to raw data sources, a thin report creates a live connection directly to the shared dataset that's already sitting in the Power BI Service. This setup dramatically simplifies the job of the report author, who can now focus entirely on telling a story with the data rather than wrestling with data modeling.

Step-by-Step Guide: How to Create and Share Your Golden Dataset

Ready to create your first reusable data model? This portion is for the person building the model - the data owner or analyst.

Step 1: Build a Flawless Data Model in Power BI Desktop

Start with a fresh Power BI Desktop file. This will be your Golden Dataset. Connect to your data sources and put on your data modeling hat. Be meticulous. This model will be used by many, so it needs to be robust.

• In Power Query: Clean your data thoroughly. Remove unnecessary columns, correct data types, handle errors, and merge or append queries as needed.
• In the Model View: Create a clean star schema with fact and dimension tables whenever possible. Ensure your relationships are correct and active.
• Add DAX Measures: Create all the essential business calculations as explicit measures. SUM(Sales[RevenueAmount]) is good, but a formal DAX measure [Total Revenue] := SUM(Sales[RevenueAmount]) is better, as it can be reused consistently.
• Document Everything: Add descriptions to your measures and tables. A user should be able to hover over a measure and a tooltip will instantly tell them what it calculates.

Leave the report canvas blank. The purpose of this PBIX file is the model, not an end-user report.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Step 2: Publish Your Dataset to the Power BI Service

Once you are confident in your model, click the Publish button on the Home ribbon in Power BI Desktop. Select a Power BI Workspace to publish to. A new, modern workspace (not "My Workspace") is required.

When you publish, Power BI automatically sends two items to the service: the (blank) report and the dataset that powers it.

Step 3: Configure "Build" Permissions for Your Team

This is the most critical step. For others to use your dataset, you must give them the right permissions.

1. Navigate to the workspace in your browser and find the dataset you just published.
2. Click the three dots (...) next to the dataset name and select Manage permissions.
3. Click Add user. Enter the email addresses, distribution lists, or security groups of the people who will be creating reports.
4. Here’s the key part: Grant them Build permission. This permission allows users to connect to your dataset from their own Power BI Desktop files, use Analyze in Excel, and create new reports on top of it, without being able to edit the original dataset itself.

You can also go an extra step and provide an endorsement - marking the dataset as "Promoted" or "Certified" to signal to others that this is an official, trusted data source.

Step-by-Step Guide: How to Connect to a Shared Dataset

Now, let’s switch hats. This part is for the report author who wants to build stunning visuals using that pre-built Golden Dataset.

Step 1: Open Power BI Desktop

Open a completely new, blank Power BI Desktop file.

Step 2: Connect to a Power BI Dataset

On the Home ribbon, instead of clicking "Get Data," click on OneLake data hub. From the dropdown menu, select Power BI datasets.

Step 3: Find and Select Your Golden Dataset

A new window will appear, showing a list of all the datasets available to you across all workspaces. You'll see any "Promoted" or "Certified" datasets at the top. Find the Golden Dataset your colleague just shared with you and click Connect.

Graphed

Still Building Reports Manually?

Watch how growth teams are getting answers in seconds — not days.

Watch the Demo

Step 4: Notice What’s Different

Once connected, your Fields pane on the right will populate with all the tables, columns, and measures from the shared dataset. It will feel just like you built it yourself, but with a few key differences:

• You cannot access the Model view or the Power Query Editor.
• You have a live connection to the dataset, indicated by a status bar at the bottom: "Connected live to the Power BI dataset: [Dataset Name]".

This is intentional! The data model is managed centrally, and you are simply a consumer of that model. Your only job is to build the best possible report with the clean data already provided to you.

Step 5: Start Building Your Visuals

You're all set! Just drag and drop fields onto the canvas to create your charts, tables, cards, and more. You can also create new report-level measures if you need a specific calculation that only applies to your report, but it’s best practice to have the core measures defined in the central model.

When you publish this "thin report," it will maintain its live connection to the shared dataset, ensuring it's always up-to-date.

Final Thoughts

Learning to use shared datasets is a major step in maturing your Power BI skills. By decoupling your data model from your report visuals, you create an efficient, scalable, and trustworthy reporting environment that saves time for both data modelers and report authors alike.

While mastering Power BI’s data modeling tools unlocks new possibilities for creating centralized datasets, we know that getting all your data connected and ready for analysis can still feel like a hurdle. We built Graphed to instantly bridge that gap. We connect directly to your marketing and sales tools - like Shopify, Google Analytics, and Salesforce - and let you create dashboards and get insights just by asking simple questions, all without wading through complex setup processes.