The 20-Year File: A Zero-Dependency Architecture for Durable Data Visualization

Publishing longitudinal data online has a quiet problem. PDFs stay accessible but go stale. Dashboards stay current but depend on frameworks, servers, and package managers that don't last. This note describes a third option, built while publishing the Michigan Cannabis Market Analysis. The approach is a single HTML file: data embedded as JSON, logic written in vanilla JavaScript, no external dependencies. The browser opens it, recalculates the summaries, redraws the figures, and gets on with it. No server wakes up. Nothing has to be installed. The file works offline, can be emailed, forked, and archived. The result is a document that stays current as the dataset grows and stays accessible regardless of what happens to the infrastructure around it. Google PageSpeed scores 100. Catchpoint records 0.0 TBT. The architecture turns out to be fast because it is simple, and durable because it travels complete.