pandas-dev/pandas
replaced: "pandas is a BSD-3 Python data-analysis library (not a service) built on NumPy with a Cython-compiled core (pandas/_libs). It provides tabular containers (DataFrame/Series on an NDFrame base), a rich label-axis model (Index/MultiIndex/DatetimeIndex/PeriodIndex/IntervalIndex/CategoricalIndex/RangeIndex), categorical and nullable extension dtypes, missing-data handling (NaN/NaT/pd.NA), split-apply-combine (GroupBy), reshaping/merging, time-series offsets/frequencies, and an unusually broad I/O surface: flat files, Excel, JSON, Parquet/Feather/ORC/Iceberg (pyarrow), HDF5 (PyTables), SAS/Stata/SPSS, SQL databases (SQLAlchemy/ADBC/sqlite3), clipboard, HTTP(S) URLs, and remote object stores (S3/GCS via fsspec). All outbound edges terminate in data systems and file backends; there are no inbound network endpoints, no auth, no message brokers, and no logging — diagnostics are delivered as warnings and exceptions."
This is the same read Preston does for a team that hires him — the difference is that a hired Preston keeps it, and reviews against it. Redeem an invite →