How to Build an Interactive Dashboard in Looker

Building an interactive dashboard in Looker turns a static report into an explorable canvas for your team. Instead of just presenting data, you can empower users to ask their own questions, filter results, and dig into the numbers that matter most to them. This guide will walk you through creating a dashboard and adding the interactive elements that make it truly powerful.

What Makes a Looker Dashboard "Interactive"?

An "interactive" dashboard is one that users can actively engage with rather than just passively view. It invites exploration. In Looker, this interactivity is driven by three main features:

• Filters: These are controls, often at the top of the dashboard, that let users narrow down the data shown in all the charts. Common examples include date range selectors, drop-down menus for marketing channels, or text boxes to search for a specific customer or region.
• Cross-Filtering: This powerful feature allows users to click on a data point or segment in one chart to automatically filter every other chart on the dashboard. For instance, clicking on "California" in a "Sales by State" map could instantly update all other charts to show data for California only.
• Drill-Downs: Drilling enables users to click on a value in a chart to see the more granular data that makes up that number. You could click on a total monthly revenue figure to see the individual transactions, or click a bar for "Organic Search" traffic to see a breakdown by landing page.

When combined, these features transform a simple report into a self-service analytics tool, reducing follow-up questions and giving your team the ability to find answers on their own.

Before You Build: Laying the Groundwork

Jumping straight into building without a plan can lead to confusing and ineffective dashboards. Before you add a single chart, it’s important to get your foundation right.

1. Ensure Your Data is Connected

Looker doesn’t store your data directly. It's a "semantic layer" that sits on top of your existing database (like BigQuery, Redshift, or Snowflake). The very first step, usually handled by a data team or an administrator, is to connect Looker to the database where your raw data lives.

Once connected, you decide which tables you want to analyze. This could be anything from your website traffic data in Google Analytics, sales records from your CRM, or order information from Shopify. Without a properly configured data connection, you won’t be able to build anything.

2. Understand the Role of LookML

Looker's magic is powered by a language called LookML (Looker Modeling Language). While you don't need to be a LookML developer to build a basic dashboard, it's helpful to know what it is. A data analyst or developer on your team uses LookML to define all your business logic. They create the "semantic model," which essentially acts as a single source of truth for your data.

Think of it this way:

• The LookML model defines what a "user," "session," "revenue," or "conversion rate" is.
• It makes dimensions (attributes, like ‘Campaign Name’ or ‘Region’) and measures (aggregations, like ‘Total Sales’ or ‘Average Order Value’) available for you to use.

So, when you drag and drop fields in Looker's interface to build a chart, it's the underlying LookML model that does the heavy lifting of writing complex SQL queries for you. If you find you're missing a metric you need, you’ll typically ask a developer to add it to the LookML model.

3. Define Your Dashboard's Purpose

This is the most critical step. A great dashboard answers specific business questions. Before starting, ask yourself and your stakeholders:

• Who is this dashboard for? (e.g., The marketing team, sales leadership, the CEO)
• What decisions will they make using this data? (e.g., Where to allocate ad spend, which sales reps need coaching, which products to promote)
• What are the 3-5 most important questions this dashboard must answer? (e.g., Which marketing channels have the best ROI? How is our sales pipeline pacing against quota? What is our customer lifetime value?)

Having clear answers will guide every decision you make, from which charts to include to which filters will be most useful.

Step-by-Step: Building Your Dashboard in Looker

Once your foundation is set, you can start building. We’ll use a basic marketing performance dashboard as an example.

Step 1: Create a New Dashboard

First, navigate to the folder where you want your dashboard to live. In the top right corner, click the New button and select Dashboard. Give your dashboard a clear, descriptive name like "Marketing Performance Overview - Q4 2023" and click Create Dashboard. You'll be taken to a blank canvas in edit mode.

Step 2: Add Visualizations (Tiles)

Dashboards are made up of individual components called "Tiles." A tile can be a chart, a table, a number, or a piece of text.

You have two primary ways to add a tile:

1. From an existing Look: A "Look" is a saved report or visualization. If you or a teammate has already built a chart you need, you can search for and add it directly.
2. From an Explore: An "Explore" is the starting point for building a new query. This is where you select your metrics and dimensions to create a new visualization from scratch.

In edit mode on your dashboard, click Add and select Visualization. This will prompt you to choose an Explore to begin building your tile.

Step 3: Build a Visualization in an Explore

Let's create our first tile: a line chart showing website sessions over time.

1. After clicking Add Visualization, select your Google Analytics Explore.
2. In the left-hand panel, you'll see a list of available Dimensions and Measures defined in your LookML model.
3. Find your date dimension (e.g., 'Traffic Date') and click on it. This will appear under 'Data'.
4. Find the measure you want to track (e.g., 'Sessions') and click on it.
5. Click Run in the top right. Looker will query your database and show you a data table.
6. Under the 'Visualization' tab, select the Line Chart icon. Looker will automatically render your data as a line chart.
7. Click the gear icon to customize the chart's colors, labels, and axes.
8. Once you're happy, give your tile a name like "Weekly Website Sessions" and click Save. It will now appear on your dashboard!

Repeat this process to add other key visualizations: a bar chart for "Sessions by Channel," a single value tile for "Total Conversions," and a table for "Top Performing Campaigns." Arrange them on the canvas by dragging and resizing the tiles.

Making It Interactive: Adding Filters and More

Now that you have your core charts, it's time to bring the dashboard to life with interactive features.

Adding Dashboard Filters

Dashboard-level filters allow users to slice and dice all the data on the page from one central place.

1. While in edit mode, click Add from the menu bar at the top, and then select Filter.
2. You'll see a pop-up window to configure your new filter. First, give it a name, like "Date Range".
3. Next, choose the Type of control. For a date range, something like "Date Range" is perfect. For a channel filter, you might use "Dropdown Menu."
4. For the "Field to Filter," select the underlying LookML field this filter should control. For our date filter, this would be the 'Traffic Date' field from our Google Analytics Explore.
5. On the next tab, "Tiles to Update," make sure you specify which tiles on your dashboard should respond to this filter. Select each tile and choose the correct field (e.g., 'Traffic Date') that the filter should apply to.
6. Click Add. Your filter will appear on the dashboard. Save your changes to exit edit mode and test it out.

Enabling Cross-Filtering

Cross-filtering is one of the easiest and most impactful features to enable.

1. Enter edit mode on your dashboard.
2. Click on Settings in the top toolbar.
3. In the 'Cross-filtering' section of the settings pop-up, you’ll likely see a checkbox to "Allow cross-filtering dashboards." Check it.
4. A list of your dashboard's tiles will appear below. Here, you can select which charts will be "selectable" (can initiate a cross-filter) and which will be "affected" (will be filtered). For most cases, you can enable both for all relevant charts.
5. Click Save.

Now, when you leave edit mode, you can click on a bar in your "Sessions by Channel" chart, and all other tiles on the page will automatically filter to show data just for that channel.

Setting Up Drill-Downs

Drill-downs give users a window into the raw data behind an aggregate number. This functionality is often configured within the LookML model itself by a developer. They create a "drill field," which specifies what happens when a user clicks on a particular measure.

From a dashboard builder's perspective, there typically isn't a setting to toggle on or off. If drill-downs have been defined in the model, they will be enabled by default. To test it, simply click on a data point in a chart (like a bar or a point on a line graph). A menu should appear, allowing you to "Show All [Records]" or drill into another, more detailed dimension. If it isn't working as expected, you’ll need to coordinate with your data team to have them configure the appropriate drill fields in the underlying model.

Final Thoughts

Building an interactive Looker dashboard is about layering filters, cross-filtering, and drill-downs on top of a well-defined set of visualizations. This transforms a static report into a dynamic tool that empowers your entire team to find their own answers and make smarter, data-driven decisions without relying on a data analyst for every ad-hoc question.

Of course, becoming proficient in tools like Looker often involves a steep learning curve and requires coordination with a technical team to get the data models right. That's why we built Graphed. We wanted to eliminate the layers of complexity and make powerful data analysis accessible to everyone. You simply connect your marketing and sales platforms in a few clicks, then ask for the dashboards and reports you need in plain English. Graphed’s AI instantly builds live, interactive dashboards, so you can go from question to insight in seconds, not hours.