What problem does it solve? Dashboards often fail to communicate because of poor layout, misleading charts, and missing context. This Skill builds and reviews dashboards according to the Japanese Digital Agency's "Dashboard Design Practical Guidebook" (March 31, 2026 edition), enforcing a 27-item checklist so every output is readable, accurate, and actionable. ## Core Features & Use Cases - Streamlit Dashboard Construction: Generates complete, runnable .py files with KPI summaries, filters, Plotly charts, the official color palette (#0017C1), and cached data loading. Supports both the latest Streamlit and Streamlit in Snowflake (SiS) 1.35 constraints. - HTML Dashboard Mode: Produces a standalone .html file using Plotly.js via CDN for users without Python or Streamlit environments. - Review Mode: Evaluates existing dashboard code against the full checklist (user experience, chart design, technical, data design, accessibility) and returns prioritized improvement suggestions with fix snippets. - Use Case: A sales manager asks for a monthly revenue dashboard. The Skill confirms requirements (5W1H), proposes a layout with the most important KPI top-left, outputs working Streamlit code with target-vs-actual comparisons, and appends a mandatory self-review table. ## Quick Start Ask the AI to build a monthly sales KPI dashboard in Streamlit following the Digital Agency design guidebook, or paste existing dashboard code and request a review.