How to Make a Scatter Plot in Looker
A scatter plot is one of the best ways to quickly see the relationship between two different variables in your data. It helps you instantly spot trends, clusters, and unusual outliers that might go unnoticed in a simple table or bar chart. This guide will walk you through exactly how to create, customize, and interpret a scatter plot in Looker (now part of Looker Studio).
What is a Scatter Plot and Why Should You Use One?
In simple terms, a scatter plot places dots on a graph to represent individual data points. Each dot has two values: one determines its position on the horizontal X-axis, and the other determines its position on the vertical Y-axis. By plotting all your data points this way, you can visually assess how the two variables relate to each other.
They’re incredibly useful for answering questions like:
“Does our marketing ad spend actually correlate with a rise in sales?”
“Are customers who spend more time on our website more likely to have a higher lifetime value?”
“Are there any outlier sales reps who are closing a high number of cheap deals compared to everyone else?”
Here are the three primary reasons to use a scatter plot:
Identifying Relationships: You can quickly see if there's a positive correlation (as one variable increases, the other tends to increase), a negative correlation (as one increases, the other tends to decrease), or no correlation at all.
Spotting Outliers: Points that fall far away from the main cluster of data are called outliers. These can represent anomalies worth investigating - like a marketing campaign that was wildly successful (or unsuccessful) compared to others.
Detecting Clusters: Sometimes data points group together in distinct clusters, which can indicate different segments or behaviors within your data. For example, you might find that customers from a certain region form a tight cluster with low spending but high purchase frequency.
Preparing Your Data in Looker
Before you build your chart, you need to make sure your data is structured correctly. Thankfully, the requirements for a scatter plot are straightforward.
At a minimum, you need two numerical fields. These can be either measures (calculations like SUM, AVERAGE, or COUNT) or numerical dimensions (like age, price, or score).
The first numerical field will be plotted on your X-axis (the horizontal line).
The second numerical field will be plotted on your Y-axis (the vertical line).
For example, if you wanted to analyze retail customer behavior, your underlying data might be an Explore based on website orders. You would want to select fields like:
Total number of orders per customer (X-axis)
Total revenue per customer (Y-axis)
Each customer would then be a single dot on the graph, and you could see if customers who place more orders also tend to generate more revenue.
Step-by-Step: How to Make a Scatter Plot in Looker
Once you know which two variables you want to compare, creating the chart itself is a simple process. Let’s walk through it together.
Step 1: Open an Explore and Select Your Fields
Navigate to the Explore where your data lives. Using the field picker on the left, select the two numerical fields you want to visualize. For this example, let's say we are a SAAS company looking at website traffic from Google Analytics. We'll select:
Users: a measure representing the total number of users.
Sessions: a measure representing the total number of sessions.
We’ll also add a dimension to serve as our data points. In this case, let's use Date, so each point on our plot will represent a specific day.
Step 2: Run Your Query
After selecting your fields, click the orange Run button in the top right. Looker will query the database and display the results in a data table below the field picker.
You'll see a table with three columns: Date, Users, and Sessions. Now, you’re ready to visualize it.
Step 3: Select the Scatter Plot Visualization
Look at the Visualization pane on the right. Looker often defaults to a line or bar chart. Click the three dots "…" to open the full menu of chart types and select the Scatterplot option.
Looker will immediately try to render a scatter plot. It might not look right at first, but don't worry - the next step is customizing it to make sense.
Step 4: Configure Your Chart Settings
This is where you tell Looker how to properly display your data. Click the Edit button in the visualization pane to open the settings menu. You’ll be working across a few tabs to dial things in.
The Plot Tab
This tab controls the fundamental structure of your chart.
X-Axis: Choose which field goes on the horizontal axis. For our example, we’ll select “Users.”
Y-Axis: Choose the field for the vertical axis. We’ll select “Sessions.” By convention, the dependent variable (the one you think is affected) often goes on the Y-axis.
Point Size: You can keep this as "Default" for uniform point size or select a third numerical field to create a bubble chart (more on that later).
Value Labels: Toggling this on will display the numerical value next to each point. This is usually too cluttered for scatter plots unless you have very few data points.
The Series Tab
Here you can control the appearance of your data points and add context with trend lines.
Colors: You can assign a custom color for your data points here.
Trend Line: This is one of the most powerful features. Click the dropdown to add a line of best fit that visually represents the overall trend. You have several types:
Linear: A straight line that works best for data that seems to follow a straight path.
Exponential: A curved line that’s useful when values increase more and more rapidly.
Logarithmic: Another curved line that's good for data that increases quickly at first and then starts to level off.
Moving Average: This smooths out short-term fluctuations to show the longer-term trend.
The X and Y Tabs
These tabs allow you to format your axes for clarity.
Scale Type: You can switch between Linear and Logarithmic scale. A log scale is useful if your data includes a wide range of values and points are getting bunched up in one corner.
Axis Name: Add a descriptive name to each axis, like "Total Website Users" and "Total Website Sessions," to make the chart easy to understand.
Format: You can change the number formatting (e.g., to currency, percentage, or rounded K/M/B).
Once you’ve configured these settings, you’ll have a clear scatter plot showing the relationship between users and sessions, with each point representing a unique day. You can now save this visualization to a dashboard.
Advanced Tips for Better Scatter Plots
Want to take your analysis a step further? Here are a few techniques to uncover deeper insights.
Add a Third Dimension with Color
A standard scatter plot shows the relationship between two variables. But what if you want to see how that relationship changes based on a third, categorical variable? Just add it to your query!
Going back to our example, let’s add the dimension Device Category (Mobile, Desktop, Tablet) to our query. With this field included, Looker provides options in the Series customization tab. You’ll be able to display each device category as a different color. You might discover that the relationship between users and sessions is much stronger on Desktop than on Mobile, an insight you'd have missed otherwise.
Use Point Size to Add Another Measure (Bubble Chart)
Just as you used color for a third dimension, you can use dot size to represent a third measure. This turns your scatter plot into a bubble chart.
For example, let's plot:
X-Axis: Marketing Campaign Spend
Y-Axis: Conversion Rate
Point Size By: Number of Leads
This chart would not only show you the relationship between spend and conversion rate, but the size of each bubble would tell you the volume of leads each campaign generated. A small bubble with a high conversion rate might be efficient, while a very large bubble shows a campaign that drove massive volume.
How to Deal with Overplotting
Overplotting happens when you have so many data points that they overlap and obscure each other, creating a dense, unintelligible blob. This is common with large datasets.
To fix this, you can try a few things in Looker’s settings:
Reduce Point Size: Smaller dots can help differentiate overlapping points.
Filter Your Data: If looking at daily data is too dense, try changing your dimension from Date to Week or Month to aggregate the data into fewer points.
Use Hexagonal Binning: available in certain chart types for mapping tools, where they represent density rather than individual points. While not a native scatter option you can create LookML to emulate this for very dense datasets.
Final Thoughts
Creating a scatter plot in Looker is an excellent technique for discovering relationships, trends, and outliers baked into your business analytics. By selecting your numerical fields, choosing the right visualization settings, and using features like trend lines and color-coding, you can turn a simple table of numbers into a powerful visual story.
While Looker provides the building blocks, getting all the settings right and pulling data from different places still requires manual work and knowing your way around the platform. At Graphed, we’ve made this process as simple as having a conversation. You can connect your marketing and sales data sources in just a few clicks and then just ask for what you need. Instead of building the chart step-by-step, you could simply type "Show me users vs. sessions from Google Analytics by device type as a scatter plot." We instantly create a live, interactive chart for you, so you can spend less time configuring reports and more time acting on the insights you find.