Colophon
Every page on this site is a static file, generated ahead of time and served as-is. There is no application server, no database behind the pages, and no code from anyone else running in your browser.
The site
Built with Astro, which renders components to plain HTML at build time and ships no JavaScript unless a page specifically needs it. Most pages here need none, so most pages send none.
Written in TypeScript under its strictest settings, with the compiler configured to reject unchecked array access, implicit type widening, and several other patterns it would otherwise allow. The trade is more time spent arguing with the type checker and less spent debugging in production.
Styled with Tailwind CSS v4, whose configuration lives in CSS custom properties rather than a JavaScript file. That matters here because the map and chart tools need to read the same palette, and custom properties are a format every one of them can already read — one source of truth instead of three copies.
Type and colour
Headings are set in Iowan Old Style where it is available, falling back to Palatino and then to Georgia. Body text uses whatever sans-serif your operating system considers default. No typeface is downloaded from this site or from anywhere else, which is partly a performance decision and partly a privacy one — a hosted font service sees every visitor who loads a page.
The palette is a warm off-white ground with a rust accent, defined in a perceptually uniform colour space so that two tones of the same lightness actually look like it. Section breaks are hard rules rather than shadows or rounded panels. The reference this is reaching for is a field guide or a well-set index, not a dashboard.
Every foreground and background pairing was checked against the WCAG contrast thresholds rather than judged by eye, and dividers that carry meaning use a darker tone than dividers that are only decoration.
Data
Every tool is built on public United States federal data. The raw sources are processed into clean, documented datasets, and those datasets are published alongside the tool that uses them — the intent is that someone can take the data and ignore the interface entirely.
Sources are cited on each tool, precisely enough that a number can be traced back to the record it came from.
What runs in your browser
Nothing from a third party. Every asset comes from this domain, and a content security policy restricts the page to exactly that — a request to an outside origin would be blocked rather than merely absent, so the guarantee does not depend on nobody making a mistake later.
There is no analytics, no tracking, and no request logging. The privacy page covers that in full.
Hosting and content delivery are handled by Amazon Web Services. The infrastructure is defined as code and deployed automatically, which is the only reason a project this small can afford to be this careful.
Source
The repository is private for now. The datasets are the part worth having, and those are published openly.
Who
One person, as a side project. There is more about that on the about page.