dashboard-design

The dashboard-design skill serves as the foundational architect for your Vizro project. It is specifically designed to guide users through a structured methodology—Requirements, Layout, and Visualization—before a single line of implementation code is written. This skill is intended for data analysts, engineers, and product designers who need to create interactive, production-grade BI tools using the Vizro Python toolkit. By mandating a rigorous planning phase, it prevents common design anti-patterns and ensures that the final dashboard is purpose-driven, user-centric, and technically sound.

• Orchestrates a mandatory 3-step workflow: Requirements gathering (Step 1), Layout & Interaction design (Step 2), and Visualization selection (Step 3).

• Produces standard YAML-based spec files (spec/1_information_architecture.yaml, spec/2_interaction_ux.yaml, spec/3_visual_design.yaml) for versioning and collaborative alignment.

• Enforces Vizro-specific design constraints including KPI card limits, navigation hierarchy, and grid-based responsive layout systems.

• Facilitates rapid decision-making by offering 2-5 numbered options for user selection, keeping the design process lightweight and iterative.

• Integrates seamlessly with follow-up skills like dashboard-build, ensuring that planning outputs directly inform the build phase.

• Uses ASCII wireframing to validate UI components, filter placement, and container interactions prior to final build.

• Always start at Step 1 if the user lacks a clear plan; do not skip directly to build implementation.

• Leverage the provided validation checklists at each step to ensure data source accessibility and actionable KPI definitions.

• For single-page or simple dashboards, steps may be abbreviated but never entirely bypassed to maintain structural consistency.

• If the user provides existing wireframes or mockups, validate existing design choices against Vizro constraints before proceeding to subsequent steps.

• Always document decisions in the spec/ directory to allow for project resumption across different AI sessions.

• Prioritize the use of built-in components like KPI cards rather than custom-coded charts unless the business use case explicitly requires high-code extensions.