pandas-dev/pandas

Pandas is the most widely used Python library for working with structured data, tables, spreadsheets, time series, and similar formats. It solves the core problem of loading, cleaning, transforming, and analyzing data that comes in rows and columns, without needing a database or specialized software. The library introduces two main data structures. A Series is a one-dimensional labeled array, similar to a single column in a spreadsheet. A DataFrame is a two-dimensional table with labeled rows and columns, similar to an Excel sheet or a SQL database table, but in memory and scriptable with Python. These structures support a wide range of operations: filtering rows by condition, grouping and aggregating data, merging multiple datasets together like SQL joins, pivoting data into summary tables, handling missing values (data gaps) gracefully, and reading or writing to formats like CSV, Excel, JSON, SQL databases, and HDF5 files. Time series analysis is a particular strength, pandas has built-in support for date ranges, frequency resampling (converting daily data to monthly, for example), moving window calculations (like rolling averages), and timezone handling. Data scientists, analysts, and engineers use pandas every day for tasks like loading a CSV of sales data and computing monthly totals, joining customer records from two different databases, cleaning up messy exported spreadsheets, or feeding processed data into machine learning models. It is typically one of the first imports in any data analysis Python script. The tech stack is Python with a core that uses NumPy (a numerical array library) for fast computation. Performance-critical internal code is written in Cython (a compiled language that extends Python) and C. Pandas is installed via pip or conda and runs on any platform where Python runs.