Dayder is a web app for finding spurious correlations in thousands of data sets. It was originally created for TerribleHack III where it just worked with 3000+ different causes of death, but we've kept improving it since then. It was made by Tristan Hume and Marc Mailhot. Since the hackathon we've expanded and optimized it to work with over 390,000 data sets at a time.Having been originally created for a "Stupid shit no-one needs and terrible ideas" hackathon, we don't even pretend that Dayder is useful. It is mostly a tech demo for making a super fast web app despite dealing with large quantities of data. Dayder uses a custom binary format, custom JS Canvas and DOM rendering, heavily optimized server-side code in Rust, caching and tuned data layouts for excellent performance. It can filter through 390,000+ data sets as you type with sub 50ms server response time and instantaneous rendering of thousands of graphs in the browser. It also can find correlations among all 390,000+ data sets in less than 2 seconds.