How to Create a Streaming Dataset in Power BI

Building a dashboard that shows what’s happening in your business right now can feel like pure magic. Real-time data streams allow you to monitor performance, catch issues, and make faster decisions without waiting for a scheduled refresh. This tutorial will walk you through how to create a streaming dataset in Power BI, turning your static reports into dynamic, live dashboards.

What is a Real-Time Streaming Dataset?

Unlike a standard Power BI report that queries a data source on a schedule (like once an hour or once a day), a real-time streaming dataset works by having data pushed into it continuously. This means your dashboard visuals can update within seconds of a new event happening, making it perfect for monitoring things like website traffic, live sales, social media mentions, or IoT sensor data.

Power BI offers three main types of real-time datasets. Understanding the difference is key to picking the right one for your goal.

• Push dataset: This is the most popular and versatile option. Data is pushed into the Power BI service and stored in an underlying database. This gives you the best of both worlds: visuals on a dashboard can update in real-time, and you can create traditional reports to analyze the historical data, build DAX measures, and use slicers. This is the one we'll focus on today.
• Streaming dataset: With a pure streaming dataset, data is also pushed to the service, but it's only stored in a temporary cache. Once the data is visualized, it’s gone. This allows for extremely low latency, as there's no database to write to. However, you can't build traditional reports or analyze its history. It's best for visualizing fast-moving data where the current value is all that matters, like a production line sensor reading.
• PubNub dataset: This is a specific type of streaming dataset that uses the PubNub streaming service. You don't need to push the data to Power BI, instead, Power BI subscribes to an existing PubNub data stream.

For most business use cases - like building a marketing performance dashboard or a live sales tracker - the Push dataset is the way to go because it provides both the live elements and the ability to look back at performance over time.

How to Set Up Your First Streaming (Push) Dataset

Let's create a Push dataset to track live website activity. We’ll build a simple model to capture a timestamp, the user’s country, the page they visited, and the revenue from their session. Don’t worry, this is easier than it sounds!

Step 1: Navigate to Your Power BI Workspace

First, log in to the Power BI Service at app.powerbi.com. You can’t create streaming datasets in the Power BI Desktop app, it must be done in the service. Navigate to the Workspace where you want your new dataset and live dashboard to reside.

Step 2: Create a New Streaming Dataset

In the top-left corner of your Workspace, click the New button and select Streaming dataset from the dropdown menu.

Step 3: Choose Your Source

Power BI will present you with three options: API, Azure Stream, and PubNub. For our purpose, we want to create a generic endpoint that we can send data to from any source. Select API and click Next.

Step 4: Define Your Data Structure

This is where you design your dataset’s table. Think of this as defining the columns you’d have in a spreadsheet.

1. Dataset name: Give your dataset a descriptive name, like "Live Website Traffic."
2. Values from Stream: Now, add the "columns" for your data stream. For our example, let's add the following:

The most important part of this step is at the bottom: toggle Historic data analysis to On. This switch is what turns our standard streaming dataset into a more powerful Push dataset, allowing Power BI to save the data it receives.

When you're happy with your structure, click Create.

Step 5: Copy Your Push URL

Once you click "Create," Power BI will show you a confirmation screen with the API info. You'll see your Push URL listed. This long, unique URL is the endpoint that data needs to be sent to. Click the copy icon to save it somewhere safe for the next step. You’ll also see a sample JSON payload, which is helpful for developers wanting to format the data correctly.

Sending Data to Your Power BI Dataset

With our streaming dataset ready, we now need to actually send some data to it. The "push" can come from anything that can make an HTTP POST request, like a custom C# application, a Python script, or a PowerShell command.

But you don't need to be a developer to get started. A user-friendly, low-code tool like Power Automate (formerly Microsoft Flow) is perfect for this. Let's build a simple flow that generates fake data and sends it to our Power BI dataset every minute.

Example: A Simple Power Automate Flow

1. Go to Power Automate and create a new Scheduled cloud flow. Set it to run every 1 minute.
2. Add a new step and search for the Power BI connector. Choose the action called Add rows to a dataset.
3. Fill in the fields:
4. Once you select the table, Power Automate will show you the fields you defined earlier (TimeStamp, UserCountry, etc.). Now you can insert dynamic content or expressions to simulate data.
5. Save and run your flow! It will now push a new row of data to your Power BI dataset every minute. The first time, it might feel like an abstract step, but behind the scenes data is now flowing into Power BI!

Visualizing Your Live Data on a Dashboard

Now for the fun part: making it visual. With real-time datasets, the primary place for building your streaming visuals is directly on a Power BI Dashboard, not in a traditional report.

1. Navigate to Your Dashboard: Go to the Power BI workspace and either open an existing dashboard or create a new one.
2. Add a Tile: In the dashboard menu, click Add a tile.
3. Choose Custom Streaming Data: From the options, select Custom Streaming Data and click Next. Your available streaming datasets will appear.
4. Select Your Dataset: Choose the "Live Website Traffic" dataset we created. The tile creation screen will now open up.
5. Build Your Visual: Now you can build your first live streaming visual.

Click Apply, and you'll see your first real-time tile on your dashboard. As your Power Automate flow continues to send data, this chart will update automatically before your eyes, with no need to hit the "Refresh" button.

Building Reports for Historical Analysis

Because we enabled "Historic data analysis," Power BI creates both your streaming dataset and a standard, reportable dataset from the same data source. With this, you can now build a traditional Power BI report with full historical data which includes all of the Power BI Desktop features you already know how to use — including sorting, filtering, and diving into the report with data drill downs. This gives you the flexibility to build a high-level operational dashboard that’s live, complemented by a detailed report for deeper dives and trend analysis.

Final Thoughts

Setting up live, streaming dashboards in Power BI is an extremely powerful way to create a centralized real-time stream of mission-critical business data into reports that anyone on your team can understand. By creating a streaming dataset via the API, sending data to its endpoint with a tool like Power Automate, and then visualizing it on a dashboard, you can build a powerful monitoring system without needing deep technical expertise.

Learning the ins and outs of tools like Power BI is incredibly rewarding, but it can sometimes feel like a heavy lift to connect all your data and start getting answers. At Graphed, we've designed an experience to remove this friction entirely, especially for busy marketing and sales teams. Instead of manually configuring data flows from sources like Google Analytics, Shopify, and Salesforce to build your dashboards, all you have to do is connect your accounts a single time. And because you are creating marketing reports and sales dashboards using natural language, your time to insight goes from hours down to just minutes. If you want to move faster from data collection to making critical business decisions, give Graphed a try and let an AI data analyst handle the busy work of creating live data streams for you.