Exploring Lakeview Dashboards for Tracking Databricks Usage

Rohit Bhagwat
4 min readMay 1, 2024

--

Note: The views expressed in this blog are my own and do not represent the official stance of Databricks.

Dashboards have long been part of Databricks SQL, originally built on the ReDash framework to provide analysts and engineers with tools for creating reports and dashboards. As someone who started my career as a BI developer, I often find myself building dashboards on whatever tools I get my hands on.

Databricks recently introduced Lakeview Dashboards, made generally available a few weeks ago. Lakeview enhances these offerings with a modern interface and advanced features, making it easier to interact with and visualize data, tightly integrated with the lakehouse architecture. While tools like Tableau and PowerBI remain relevant, Lakeview presents a powerful, integrated alternative for SQL-savvy users creating shared analytics.

Refer to the official blog to learn more

Let’s explore some of the new capabilities of Lakeview dashboard, we’ll use system tables to explore historical data across the Databricks account.

At the heart of Lakeview Dashboards are two fundamental components: Canvas and Data.

Canvas & Visualizations

The single-page canvas layout lets users add visualizations, apply filters, and incorporate markup-supported text. The refreshed visualization toolkit is a pleasant update.

The ‘Text to Visualization’ feature offers a neat trick to use natural language to create visualizations.

Data Management

Lakeview’s object model supports dynamic, parameterized queries, allowing users to select from existing tables or craft custom queries. This system is optimized for performance, handling smaller datasets client-side for quick interactions, while larger datasets utilize SQL compute for intensive processing.

The integration of Databricks Assistant within the Data tab significantly enhances the SQL query writing experience in Lakeview Dashboards. Leveraging Databricks AI and Unity Catalog, the Assistant understands the semantic layer of the lakehouse deeply.

This means you can pose straightforward questions like, ‘help me write a query to analyze DBU usage by month across SKUs’ and the Assistant will craft the SQL.

Sharing

I’m glad that now we can publish the dashboards and share with anyone in the enterprise as long as they have a Databricks account (not necessarily on the same workspace). This includes additional capabilities for scheduled exports and PDF generation through Databricks Workflows. I look forward to being able to be able to embed these dashboards in upcoming releases.

An important addition is the ability to manage versions, as you develop and refine your dashboards, Lakeview allows for both a published version and a draft version. This offers the flexibility to experiment and make changes without affecting the end users’ experience.

API Integration

The API layer of Lakeview is designed to support automation in deploying and managing dashboards, which is essential for integrating these tools into larger CI/CD pipelines. The APIs include the ability to publish, unpublish, create, update or even migrate a dashboard.

API Documentation

Usage Dashboard

I invite you to download and explore the usage dashboard I’ve developed while experimenting with Lakeview Dashboards. You can see it in action within your own environment — just ensure you have access to system tables.

https://github.com/rohit-db/public

Migrating from Legacy Dashboards

Migrating from Legacy dashboards to Lakeview is streamlined to ensure a smooth transition. When you clone an existing dashboard to a Lakeview dashboard, Databricks automatically transfers the queries and sets up the datasets in the Data tab. However, be aware that some specific functionalities from Legacy dashboards might not yet be supported in Lakeview.

Closing Thoughts

Lakeview dashboards introduce a sleek new interface and essential functionalities like Text to Visualization, making them accessible for both seasoned SQL users and newcomers to the Databricks platform.

I’m particularly excited about future updates, including multi-page support and expanded visualization options. These anticipated features will undoubtedly broaden Lakeview’s adoption and enhance user experience even further.

Try Lakeview dashboards and share your thoughts here!

--

--

Rohit Bhagwat

Data & Analytics Professional, Aspiring Data Scientist, Runner, Gadget enthusiast