Datascience is a curated list of Python resources for data science that aggregates libraries, tutorials, code snippets, blog posts, and talks.
The project addresses the challenge of discovering and organizing the fragmented ecosystem of Python data science tools. Rather than building new functionality, it catalogs existing resources across the full workflow: core libraries like pandas, scikit-learn, and matplotlib for data manipulation and visualization; general Python programming tools for dependency management, logging, and configuration; and alternatives to pandas such as polars, duckdb, and daft for specialized use cases. The curation approach helps practitioners navigate options without needing to search multiple sources.
Someone considering this resource should understand it serves as a reference guide rather than a learning platform or framework. It suits developers building data science projects who need to evaluate tool choices or discover libraries they may have missed. The list is particularly valuable for understanding the landscape of pandas alternatives, which it organizes by use case: polars and fireducks for performance, duckdb for SQL integration, xarray for multidimensional data, and modin or swifter for parallelization. The resource does not position itself against other curated lists but instead focuses on breadth across the Python data science stack.
The project maintains an organized structure with clear categorization of resources and includes descriptive context for many entries, such as noting that ruff replaces flake8, black, and isort, or that narwhals enables dataframe-agnostic code across multiple libraries. Updates appear to reflect the evolving Python ecosystem, with recent additions of tools like uv for dependency management and fireducks as a pandas alternative.
