UCI ML Repository

archive.ics.uci.edu

machine learning के benchmark datasets का पुराना archive। हर dataset पेज variables और provenance को दस्तावेज़ करता है — raw CSV नहीं, पेज import करें।

Each dataset page documents the variables, collection method, and known issues, which is what you want in a notebook explaining a classic dataset like Iris or Adult Income rather than just handing over numbers with no context. It fits teaching notebooks that walk through a specific algorithm or feature-engineering problem. Most of the datasets are small, tabular, and decades old at this point, so it won't help with anything needing modern-scale or unstructured data.