Short, practical write-ups on operational data work: metric audits, GitOps for BI, Superset in production, and the platforms we've actually shipped.
documented in GitHub issues, unfixable by configuration
Theme flashes, SSO session eviction, dead third-party cookies, trapped filter state, 15-second loads, and exposed admin links — the six structural pain points of the official Superset embedding SDK, with receipts.
guest tokens, RLS, and the config nobody documents in one place
A walkthrough of the official @superset-ui/embedded-sdk: the iframe lifecycle, the guest token flow, row-level security, and every config switch you need to flip.
what actually works for large deployments
Why Apache Superset's flat dashboard list breaks at fifty assets, why the native TAGGING_SYSTEM introduces database and UX failures, and how curated folder hierarchies scale to hundreds of dashboards.
an engineering-first comparison of the two open-source BI platforms engineers actually ship
Stack and runtime model, SQL-first vs visual-first, where the governance paywall lands, and what embedded licensing really costs — a side-by-side for engineers picking an open-source BI stack, not a marketing funnel.
an engineering-first comparison for teams picking a BI stack inside — or outside — the Microsoft orbit
Where queries run, how Fabric capacity pricing actually behaves, what Entra ID lock-in means in practice, and how embedded licensing diverges — a side-by-side for engineers picking a BI stack, not a marketing funnel.
an engineering-first comparison for teams that have to live with the choice
Where queries run, how the semantic layer behaves, what governance actually costs, and how embedded licensing differs — a side-by-side for engineers picking a BI stack, not a marketing funnel.
certifying datasets via GitHub Actions
How we treat Superset datasets as code: a GitHub-as-authority workflow with schema validation, Jinja parsing, SQLFluff, and a certification badge that points back to the exact CI run.