Video snapshot: what the walkthrough covers

In a new YouTube video, Pragmatic Works presenter Zane Goodman offers a beginner-friendly tour of Databricks AI/BI Dashboards. He shows how users can move from SQL queries to interactive, shareable dashboards without leaving the Databricks workspace. Moreover, the video emphasizes a simple mental model: dashboards are driven by datasets, and that idea shapes how visuals, filters, and interactions behave.

Goodman walks viewers through creating datasets from SQL or by selecting tables managed in Unity Catalog, and he demonstrates tools that speed up common tasks. For example, he uses Genie to rewrite SQL for quick adjustments and to help generate queries from natural language prompts. The presentation aims to be practical, especially for analysts who still export query results to spreadsheets to make charts.

How datasets and queries power visuals

The video makes it clear that the dataset is the central unit in this workflow: visuals read results from datasets rather than from isolated charts. Consequently, you create a dataset either by writing SQL or by selecting a governed table, and the dataset result then feeds the widgets on the canvas. As a result, changes to the dataset propagate to visualizations, which simplifies maintenance when multiple charts depend on the same source.

Furthermore, Goodman highlights how Genie helps when you need to modify queries quickly, reducing the need for hand-editing complex SQL. He also contrasts two approaches for derived fields: using post-run custom calculations or defining parameters up front, noting that each approach has its place. Therefore, teams should consider how often values change and who needs to edit definitions when choosing between the two.

Building the canvas and configuring widgets

After establishing datasets, the video moves to the dashboard canvas where widgets and layouts take shape. Goodman demonstrates drag-and-drop placement, widget sizing, and the basic widget configuration options that control labels, axes, and data bindings. He also shows how markdown or text blocks can add context, which improves readability for business users who rely on narrative alongside charts.

Importantly, he points out that thoughtful widget configuration reduces later rewrites, especially when dashboards are reused across teams. In addition, he recommends testing visuals with real data and with different filter settings so that axis ranges and aggregations behave predictably. This practical approach reduces surprises when dashboards go into production.

Filter behavior: page-level versus global

Goodman devotes time to explaining filter scope, a key interaction topic that affects user experience across pages. He distinguishes page-level filters that affect only a single dashboard page from global filters that propagate across multiple pages, and he warns about accidental cross-page filtering. As a result, designers must plan filter scope deliberately to avoid confusing users who expect separate views per page.

Moreover, he recommends setting clear default states for global filters and documenting when filter changes will affect other pages. This guidance helps collaboration, since users who share dashboards will understand which filters are persistent. Ultimately, the right balance between global consistency and page-specific detail depends on use cases and audience needs.

Publishing, sharing, and permission tradeoffs

The video closes by walking through publishing options and the tradeoffs involved in data permissions and caching. Goodman contrasts a shared data permission model that enables caching for performance with an individual permission model that ensures each viewer sees only their authorized rows. Therefore, teams must weigh performance gains against strict access requirements when they choose a publishing strategy.

Furthermore, he notes that shared caching improves responsiveness for large audiences but requires careful governance to avoid data leaks. Conversely, enforcing individual permissions increases query volume and can raise cost and latency, which creates an operational challenge for high-concurrency dashboards. Consequently, organizations should test expected loads and monitor both performance and compliance metrics.

Practical takeaways and implementation challenges

Overall, the video delivers concrete steps for teams adopting Databricks AI/BI Dashboards while acknowledging tradeoffs between ease of use and governance. For example, low-code features and Genie accelerate authoring, but teams still need clear standards for dataset design, naming, and permissioning to scale safely. In addition, balancing real-time refresh needs with cost control remains a common operational concern.

In practice, start small with clear dataset ownership and documented filter behavior, then incrementally expand dashboard scope as requirements stabilize. Finally, Goodman’s walkthrough serves as a useful primer: it equips analysts with practical techniques and warns decision-makers about the performance, security, and governance choices they must manage. As a result, organizations can adopt these dashboards more confidently and with fewer surprises.