The Python library commonly used for data manipulation and analysis with DataFrame objects is pandas.
Pandas provides two central labeled data structures: Series for one-dimensional data and DataFrame for two-dimensional, table-like data. A DataFrame can hold columns with different data types, use meaningful row and column labels, and support operations such as filtering, grouping, joining, reshaping, sorting, and handling missing values.
The project began at AQR Capital Management in 2008 and was open-sourced in 2009. Its name combines “panel data,” a term from econometrics, with a common Python naming style. Pandas is widely used with NumPy and integrates with tools such as Matplotlib, Jupyter, SQL databases, and Excel files.
The alternatives in the question have different primary roles. NumPy supplies fast numerical arrays, Matplotlib creates visualizations, and scikit-learn focuses on machine learning. They often appear in the same workflow, but pandas is the library most specifically associated with labeled tabular data and DataFrame-based analysis. It is commonly imported with the abbreviation `pd`.