How to Make a Clustered Column Chart in Looker with AI

Creating a clustered column chart in Looker is a fantastic way to compare different data categories side-by-side. Instead of wading through tables of numbers, you get an instant visual snapshot of performance. This article will walk you through what these charts are, why they're so effective, and exactly how to build one in Looker - both the traditional way and using modern AI features to do it in seconds.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

What is a Clustered Column Chart?

Think of a standard bar chart that shows sales for different products. Now, imagine you want to see the sales for those same products but broken down by two different regions, like North America and Europe, for a direct comparison. A clustered column chart - also known as a grouped column chart - lets you do exactly that.

It groups multiple columns together for each category on your main axis. In our example, the "Product" would be on the x-axis, and for each product, you'd have two columns clustered together: one for North American sales and one for European sales. This makes it incredibly easy to see which region sold more of a particular product and to compare overall performance patterns across your entire product line.

Unlike a stacked column chart which shows you the parts of a whole, a clustered chart places the bars next to each other, making it ideal for direct comparison between sub-groups.

For example, you could compare:

• Website sessions from different traffic sources (like Google, Facebook, and Email) month-over-month.
• New vs. returning customer revenue for each product category.
• Key metrics like clicks, impressions, and conversions for different ad campaigns.

Why Use Clustered Column Charts for Your Business?

Visualizing data this way isn't just about making things look nice, it's about uncovering insights you might otherwise miss. Manual reporting often involves dumping data into a spreadsheet and trying to make sense of endless rows. This process of downloading CSVs on a Monday, wrangling the data, building a report for a Tuesday meeting, and then answering follow-up questions on Wednesday eats up half your week.

Clustered column charts help you skip straight to the insight. Here’s why they are so effective:

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

1. Direct, At-a-Glance Comparisons

The primary strength of a clustered column chart is its ability to facilitate immediate comparison. When bars are side-by-side, your eyes can quickly measure differences in height an order of magnitude faster than comparing numbers in a table.

Relatable Example: A sales manager wants to review team performance for the last quarter. With a clustered chart, they can see each sales rep's progress towards their "New Deals," "Closed-Won Deals," and "Deal Value" targets clustered together. They can instantly spot who is great at opening doors but struggling to close, or who brings in the biggest deals.

2. Identify Trends and Patterns Within Sub-Groups

By grouping your data, you can uncover relationships that aren't obvious in a simple bar chart. You might discover that while one campaign has high traffic, another has a higher conversion rate - an insight that helps you reallocate your budget more effectively.

Relatable Example: A digital marketer is analyzing performance for a new product launch. They create a clustered column chart comparing website traffic from Instagram and TikTok for Week 1 vs. Week 2. They might see that while TikTok drove a massive spike in traffic in Week 1, the Instagram traffic was more sustained and grew in Week 2, suggesting a difference in the audience's buying cycle between platforms.

3. Communicate Complex Data Clearly

These charts are intuitive and easy for any stakeholder to understand, regardless of their data literacy. You don’t need to be a data analyst to grasp the story being told when you see columns of different heights side-by-side. This makes your reports more impactful and helps drive quicker, more confident decisions across the team.

How to Manually Create a Clustered Column Chart in Looker

While Looker is a powerful BI tool, learning its interface can take time. If you've ever felt like you need to take an 80-hour course just to feel proficient, you're not alone. Here are the manual steps to create a clustered column chart, which involves navigating its interface and knowing which settings to tweak.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

Step 1: Start in an Explore

Your journey begins in the "Explore" section of Looker. This is where you access your data models. Navigate to the Explore that contains the data you want to visualize. For instance, if you're analyzing web traffic, you might start in a Google Analytics Explore.

Step 2: Select Your Dimensions and Measures

In the field picker on the left, you'll need to select your data points.

• Dimensions: These are your qualitative, categorical data points - the "what" or "who." You will typically need at least two dimensions to pivot for a clustered column chart. For example, if you want to compare traffic by source over time, your dimensions could be Traffic Source and Created Date (by month).
• Measures: These are your quantitative, numerical data points - the "how many" or "how much." An example measure would be Sessions or Pageviews.

Once selected, click the Run button to pull the data into a table.

Step 3: Pivot by a Dimension

To create the "clusters," you need to pivot one of your dimensions. This transforms the rows of that dimension's data into columns. To do this, find the dimension you want to group your bars by (e.g., Traffic Source), click the gear icon next to its name, and select Pivot.

Run the query again. You'll see your data table has changed, with a new column for each unique value in your pivoted dimension (e.g., a column for "Google," one for "Facebook," etc.).

Step 4: Go to the Visualization Tab

With your data correctly formatted, it's time to build the chart. Click on the Visualization tab above your data table.

Step 5: Choose and Configure the Column Chart

Looker provides several chart types. Click the Column chart option (it looks like a bar chart icon).

Looker will automatically attempt to generate a chart. It will likely default to a stacked column chart first. To change it to a clustered chart:

1. Click the Edit button to open the visualization settings panel.
2. Go to the Plot tab.
3. Find the setting called Series Positioning.
4. Click the dropdown menu and change it from "Stacked" or "Stacked Percentage" to Grouped.

Your chart will instantly transform into a clustered column chart, with your pivoted dimension creating the groups. From here, you can use the other settings tabs (X-Axis, Y-Axis, Series) to customize labels, colors, and titles to make your chart presentation-ready.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

The Faster Alternative: Using AI to Build Your Chart

The manual process above is effective, but it requires you to know where every setting is. This process has layers: you need to understand the underlying data, you need to learn the software, and then you need to perform the right sequence of clicks and configurations to get the output you want.

The modern approach abstracts all that away with AI. Newer versions of tools like Looker, and other modern analytics platforms, are integrating Natural Language Processing (NLP). This means you can simply describe the chart you want in plain English, and the AI builds it for you.

Instead of clicking through menus, you can just type a prompt like:

"Show me website sessions by traffic source for the last 90 days. Break it down by device type. Put it into a clustered column chart."

The AI assistant understands the intent behind your words:

• It identifies "sessions" as the measure.
• It recognizes "traffic source" and "device type" as dimensions.
• It parses "last 90 days" to set the correct time filter.
• It knows a "clustered column chart" requires pivoting by one of the dimensions and setting the series positioning to "Grouped."

This allows you to create the exact visualization you need without having to touch a single configuration menu. It turns a multi-step, multi-click process into a single sentence.

Tips for Writing Productive AI Prompts

Getting great results from an AI assistant is about providing clear instructions. You don’t have to be perfect, as the AI is designed to interpret even simple prompts, but here are a few tips to get exactly what you need:

• Be Specific About the Chart Type: Start by stating your goal. "Create a clustered column chart..." or "show me a bar graph of..."
• Clearly Name Your Metrics and Dimensions: Use the terminology from your data source if you know it (e.g., "new users," "deal stage," "ad spend"). If you don't, often just a general term like "website traffic" is enough for the AI to infer what you mean.
• Define Your Timeframe: Always include a time range like "this quarter," "last 6 months," or "from January 1 to March 31." Lack of a time frame is a common source of confusion.
• Iterate and Refine: Start with a simple request ("Show me revenue by product category"). Once you have that, you can add layers of complexity in a follow-up prompt ("Now, cluster that by new vs. returning customers"). This allows you to explore your data conversationally, drilling down deeper as new questions arise from the last visualization.

Final Thoughts

Knowing how to build a clustered column chart is an essential skill for anyone looking to make sense of their business data. Whether you follow the manual steps in a traditional BI tool or leverage the speed of an AI assistant, the goal is the same: to turn raw numbers into clear, comparative insights that help you make better decisions.

At Graphed, we've built our entire tool around the belief that getting those insights shouldn't be a chore. Instead of forcing you to learn a complex piece of software, we connect to all your data sources - from Google Analytics and Shopify to your ads platforms and CRM. You can then ask questions in plain English, and our AI instantly creates the live dashboards and charts you need, without you ever having to worry about pivoting data or finding the right visualization settings. This approach turns hours of reporting busywork into a five-second conversation, empowering your entire team to find the answers they need.