How to Create a Donut Chart in Looker with AI

A donut chart is the perfect way to show how different parts make up a whole, like how much of your website traffic comes from each marketing channel. This tutorial will walk you through creating a compelling donut chart in Looker Studio (formerly Google Data Studio), first using the standard navigation and then showing you how a modern AI approach can get you there in a fraction of the time.

What Exactly is a Donut Chart?

Think of a donut chart as a pie chart with the middle cut out. That empty space isn't just for looks, it's a great spot to display a key total metric, like "Total Sales" or "Total Users." This gives viewers immediate context for the proportions they're seeing in the surrounding ring.

Donut charts are excellent for answering questions like:

What are our top-selling product categories as a percentage of total revenue?
Which social media platforms drive the most traffic to our website?
How are our marketing budget allocations distributed across different campaigns?

They work best when you're comparing a smaller number of categories - ideally, seven or fewer. Any more than that, and the slices become too small to read effectively, making the chart cluttered and confusing. In those cases, a bar chart often works better.

Creating a Donut Chart in Looker Studio (The Standard Way)

Let's build a classic donut chart to visualize website traffic sources from a Google Analytics 4 property. This process involves manually adding a chart and configuring its data fields through the Looker Studio interface.

Step 1: Connect Your Data Source

Before you can visualize anything, you need data. For this example, we'll assume you're connecting a Google Analytics account.

Open a new Looker Studio report.
You'll be prompted to connect to data. Search for and select Google Analytics.
Authorize Looker Studio to access your Google account if you haven't already.
Select the Account and GA4 Property you want to pull data from. Then click Add.

Your data source is now linked to the report, and you can start building visualizations.

Step 2: Add a Chart to Your Report

With your data source ready, you can now add the donut chart to your canvas.

In the top menu bar, select Add a chart.
In the dropdown menu, find the Pie section.
Select the second icon, which is the Donut chart.
Click anywhere on your report canvas to place the chart. Looker Studio will generate a default chart based on dimensions and metrics it thinks are most relevant. Now we'll customize it.

Step 3: Configure Your Chart's Dimensions and Metrics

This is where you tell Looker Studio what data to actually show. You do this in the Setup panel that appears on the right when your chart is selected.

Every chart needs at least two things:

Dimension: This is the categorical data you want to group by - the "slices" of your donut.
Metric: This is the numerical data you want to measure for each slice. It determines the size of each slice.

To see sessions by channel, we'll configure it like this:

In the Dimension field, search for and select Session default channel group. This will define our slices as "Organic Search," "Direct," "Paid Search," etc.
In the Metric field, search for and select Sessions. This will measure the total number of sessions for each channel group, determining the size of the slice.

Your chart will instantly update to display your website sessions broken down by channel group. Easy, right? Now let’s make it look even better.

Step 4: Style and Customize Your Donut Chart

A good chart is easy to read. In the property panel on the right, switch from the Setup tab to the Style tab to improve its appearance.

Here are key customizations:

Colors: You can change the slice colors manually to match your company branding or manage colors by Dimension value to assign specific colors to specific channels (such as making Organic Search always blue).
Labels: Adjust what's displayed to show not only the percentage but also the text.
Legend: Adjust the positioning and font family.
Hole Size: Increase or decrease the inner radius to display the number summary.

After styling, your donut chart is complete. It's effective on its own and a critical component in a larger website performance dashboard that provides a great "at-a-glance" view before stakeholders dig deeper.

Beyond One Chart: The Challenge of the Full Reporting Workflow

The manual process in Looker is perfectly functional, but it highlights a common friction point in data analysis. Building a single chart is one thing, constructing an entire dashboard full of them and answering follow-up questions quickly is another.

Every marketing or sales team knows the routine: someone asks a question like "Which marketing channels brought in the most revenue last quarter?" To answer this, you might have to:

Log into Google Analytics (and maybe Salesforce or Google Ads, too).
Export a few different CSV reports.
Spend 20 minutes in Excel PivotTables cleaning and combining the data.
Finally, build a chart to visualize the result.

Someone inevitably looks at the chart and asks a follow-up: "Okay, but what was the return on ad spend for just the search channel?" and the entire process starts over.

Building charts manually, whether in Looker or another tool, requires you to know exactly where to click and what specific database field name to select - remembering things like "sessiondefaultchannel" versus "Traffic." It's not difficult, but it does create a small barrier between having a question and knowing how to turn that into data that can answer it.

How AI Simplifies the Entire Process

Instead of manually constructing reports by clicking through a menu system, AI-based analytics tools allow you to go straight to asking your main question plainly - the same way you might ask a colleague:

“Build a donut chart that shows the number of weekly sessions by session channel for last month.”

An AI data analyst can parse that sentence to understand a "donut chart" needs both dimensional categorical data like "session default", as well as the metric "(sessions in Google Analytics)".

The Power of Asking Plainly

AI changes this dynamic radically by eliminating the interface learning curve:

No Technical Knowledge Needed: People no longer need to learn technical definitions like "dimension" or "metric." Just plain simple English.
Instant Answers: The chart you asked for is built directly, so you don't spend time configuring chart settings, it saves minutes every time you analyze - and hours when building out a complete dashboard from scratch.
Deeper, Faster Exploration: Follow-up questions like "what's the ROS from the Search channel" now take seconds instead of another 10-15 minutes of digging. AI data analysts encourage truly interactive data exploration: each quick answer prompts more questions, leading you to find real, deeper insights because the work is no longer so intensive or slow - it allows us to go deeper!

Rather than stopping analysis to set up different charts and filters, you remain focused on the story of what your business numbers really mean so that you can move quickly on making better business decisions now and acting upon discovered insights, rather than waiting hours - or even days. The result is your team spending much less time doing 'data entry,' with more resources acting more strategically and creatively.

Final Thoughts

Creating a beautiful, effective donut chart in Looker Studio is a must to better visualize proportional data - whether it be showing how marketing revenue segments, channels are split, understanding your budget spend, or sales contributions per month. It offers clear, insightful visuals if you're comfortable navigating their menus.

When you start to get tired from spending hours in Looker Studio trying to perfectly align the boxes on your reports - we built Graphed. Instead of needing to click through all the menus to build each dashboard completely, just use simple natural language and describe what you want, like "revenue vs. Ad Spends from Facebook for Q3." AI will create it in real time - then just watch how much you enjoy being freed to do so much for free! The days of spending hours fighting Looker Studio reporting issues are done.