Welcome to DataHaskell (www.datahaskell.org)

dataHaskell is being relaunched with a clear, pragmatic promise: make data science and machine learning in Haskell welcoming, practical, and fast. The team has an ambitious two-year roadmap and is focusing on a single “happy path” that solves common DS/ML workflows so newcomers can get a notebook, load data, and run a model in minutes. The reboot emphasizes people-first community norms—friendly on-ramps, predictable cadence (monthly community calls, fortnightly help-wanted sweeps, monthly release notes), and explicit ownership with named maintainers and contributor guides to build continuity and trust. Technically, the project will prioritize developer experience and reproducibility: runnable, diataxis-style documentation to prevent drift; reduced time-to-first-run via documented environments and maintainer-driven fixes; and an opinionated, narrow stack that favors one setup that “just works” over many interchangeable tools. Short-term technical goals include better plotting, faster I/O, GPU ergonomics, clearer error messages, and more copy-paste examples. The team asks users to try the stack, report friction, file focused PRs, or help polish docs and tutorials—small contributions are the main accelerator. Join the Discord to test, give feedback, or contribute as dataHaskell iterates toward a smoother Haskell data ecosystem.