{
  "aliases": [],
  "workflows": [
    "Clean and reshape survey or market data in a notebook",
    "Build features for scikit-learn or statsmodels models",
    "Load SQL query results for exploratory analysis",
    "Resample intraday prices to daily bars"
  ],
  "advantages": [
    "Largest ecosystem and documentation base in Python data work",
    "Extensive I/O coverage including statistical package formats",
    "Rich time-series tooling",
    "BSD license"
  ],
  "limitations": [
    "Single-threaded eager execution",
    "In-memory only; large datasets need other engines",
    "Multiple overlapping idioms and copy/view pitfalls",
    "Object dtype columns are slow and loosely typed"
  ],
  "prerequisites": [
    "Python 3 with NumPy; PyArrow recommended for Parquet and Arrow dtypes"
  ],
  "audiences": [
    "Data analysts and scientists",
    "Economists and finance researchers",
    "Students learning data analysis in Python"
  ],
  "alternatives": [
    "polars",
    "duckdb",
    "numpy",
    "dask",
    "r"
  ],
  "tags": [
    "dataframes",
    "python",
    "time-series",
    "open-source",
    "econometrics"
  ],
  "use_cases": [
    "analyze-intraday-data",
    "transform-data",
    "estimate-causal-effects"
  ],
  "operating_systems": [
    "macos",
    "linux",
    "windows"
  ],
  "deployment_models": [
    "embedded"
  ],
  "integrations": [
    "NumPy",
    "Apache Arrow",
    "matplotlib",
    "scikit-learn",
    "statsmodels",
    "Jupyter",
    "SQLAlchemy",
    "Polars",
    "DuckDB"
  ],
  "data_formats": [
    "CSV",
    "Excel",
    "JSON",
    "Parquet",
    "ORC",
    "HDF5",
    "SQL",
    "Stata",
    "SAS",
    "SPSS",
    "Feather"
  ],
  "languages": [
    "Python",
    "Cython",
    "C"
  ],
  "self_hostable": true,
  "discovery_method": "editorial-research",
  "depth": "deep",
  "editorial_status": "reviewed",
  "visibility": "public",
  "featured": false,
  "author": "batch-D",
  "slug": "pandas",
  "name": "pandas",
  "vendor": {
    "name": "pandas development team (NumFOCUS)",
    "slug": "pandas-dev",
    "website": "https://pandas.pydata.org/"
  },
  "homepage_url": "https://pandas.pydata.org/",
  "docs_url": "https://pandas.pydata.org/docs/",
  "repo_url": "https://github.com/pandas-dev/pandas",
  "license_url": "https://github.com/pandas-dev/pandas/blob/main/LICENSE",
  "product_type": "library",
  "product_status": "active",
  "summary": "The standard Python DataFrame library for labeled tabular data, time series and I/O to CSV, Excel, SQL, Parquet and more",
  "description": "pandas is the long-established Python library for tabular and time-series data, developed by an open community under the NumFOCUS umbrella and released under the BSD 3-Clause license. Its DataFrame and Series objects carry a labeled index, which drives alignment in arithmetic, joins, reshaping and time-based resampling. Readers and writers cover CSV, Excel, JSON, Parquet, ORC, HDF5, SQL databases, Stata, SAS and SPSS files, and the groupby, pivot, merge and window machinery handles most everyday analysis. Since version 2.0 columns can be backed by Apache Arrow arrays in addition to NumPy, improving string handling and interchange with PyArrow, Polars and DuckDB; the 3.0 line continues to move default string storage and copy semantics in that direction.\n\npandas is the lingua franca of the Python data ecosystem: scikit-learn, statsmodels, matplotlib, seaborn, Plotly and most data APIs accept or return DataFrames, and nearly every tutorial assumes it. Economists and financial researchers rely on its DatetimeIndex, business-day offsets, resampling and rolling statistics.\n\nThe library executes eagerly on a single thread and keeps data in memory, so performance and memory use become limiting on tens of millions of rows; Polars, DuckDB or Dask are the usual escapes. Its API accumulated several ways to do the same thing (loc, iloc, chained indexing, copy-versus-view ambiguities), which is a frequent source of bugs that recent versions address with copy-on-write. Mixed-type object columns and legacy NaN handling also complicate strict typing.",
  "capabilities": [
    "Labeled DataFrame and Series with index alignment",
    "Readers and writers for CSV, Excel, JSON, Parquet, SQL, HDF5, Stata and more",
    "GroupBy, pivot, merge, reshape and window operations",
    "Time-series resampling, offsets and rolling statistics",
    "Arrow-backed and NumPy-backed column dtypes",
    "Missing-data handling and categorical types",
    "Integration with matplotlib plotting"
  ],
  "comparison_notes": "Polars offers a multithreaded, lazily optimized alternative with a different expression API and no index. DuckDB brings SQL and out-of-core execution to the same files and converts to pandas when needed. NumPy underlies pandas and is preferable for homogeneous numeric arrays. Dask parallelizes a pandas-like API across cores or machines. R's data.frame and tidyverse are the closest equivalents outside Python.",
  "categories": [
    {
      "slug": "dataframes-arrays",
      "primary": true
    },
    {
      "slug": "math-stats-tools"
    }
  ],
  "execution_modes": [
    "local"
  ],
  "interfaces": [
    "library"
  ],
  "source_model": "open-source",
  "license_spdx": "BSD-3-Clause",
  "pricing_model": [
    "open-source",
    "free"
  ],
  "sources": [
    {
      "url": "https://pandas.pydata.org/",
      "type": "official_site",
      "supports": [
        "name",
        "summary"
      ],
      "checked_at": "2026-11-02",
      "title": "pandas – Python Data Analysis Library"
    },
    {
      "url": "https://pandas.pydata.org/docs/",
      "type": "official_docs",
      "supports": [
        "capabilities",
        "data_formats"
      ],
      "checked_at": "2026-11-02",
      "title": "pandas documentation"
    },
    {
      "url": "https://github.com/pandas-dev/pandas/blob/main/LICENSE",
      "type": "license_page",
      "supports": [
        "license_spdx",
        "source_model"
      ],
      "checked_at": "2026-11-02",
      "title": "pandas LICENSE (BSD 3-Clause)"
    },
    {
      "url": "https://github.com/pandas-dev/pandas",
      "type": "repository",
      "supports": [
        "repo_url"
      ],
      "checked_at": "2026-11-02",
      "title": "pandas-dev/pandas"
    }
  ],
  "confidence": "high"
}