You are reading an old version of the documentation (v2.2.4). For the latest version see https://matplotlib.org/stable/users/installing.html
Version 2.2.4
matplotlib
Fork me on GitHub

Table of Contents

Related Topics

Installing

Note

If you wish to contribute to the project, it's recommended you install the latest development version.

Installing an official release

Matplotlib and most of its dependencies are all available as wheel packages for macOS, Windows and Linux distributions:

python -mpip install -U pip
python -mpip install -U matplotlib

Note

The following backends work out of the box: Agg, ps, pdf, svg and TkAgg.

For support of other GUI frameworks, LaTeX rendering, saving animations and a larger selection of file formats, you may need to install additional dependencies.

Although not required, we suggest also installing IPython for interactive use. To easily install a complete Scientific Python stack, see Scientific Python Distributions below.

Windows

In case Python 2.7 or 3.4 are not installed for all users, the Microsoft Visual C++ 2008 (64 bit or 32 bit for Python 2.7) or Microsoft Visual C++ 2010 (64 bit or 32 bit for Python 3.4) redistributable packages need to be installed.

macOS

If you are using Python 2.7 on a Mac you may need to do:

xcode-select --install

so that subprocess32, a dependency, may be compiled.

To use the native OSX backend you will need a framework build build of Python.

Linux

On extremely old versions of Linux and Python 2.7 you may need to install the master version of subprocess32 (see comments).

Test Data

The wheels (*.whl) on the PyPI download page do not contain test data or example code. If you want to try the many demos that come in the Matplotlib source distribution, download the *.tar.gz file and look in the examples subdirectory. To run the test suite:

  • extract the lib\matplotlib\tests or lib\mpl_toolkits\tests directories from the source distribution;
  • install test dependencies: pytest, mock, Pillow, MiKTeX, GhostScript, ffmpeg, avconv, ImageMagick, and Inkscape;
  • run py.test path\to\tests\directory.

Third-party distributions of Matplotlib

Scientific Python Distributions

Anaconda and Canopy and ActiveState are excellent choices that "just work" out of the box for Windows, macOS and common Linux platforms. WinPython is an option for windows users. All of these distributions include Matplotlib and lots of other useful (data) science tools.

Linux : using your package manager

If you are on Linux, you might prefer to use your package manager. Matplotlib is packaged for almost every major Linux distribution.

  • Debian / Ubuntu: sudo apt-get install python3-matplotlib
  • Fedora: sudo dnf install python3-matplotlib
  • Red Hat: sudo yum install python3-matplotlib
  • Arch: sudo pacman -S python-matplotlib

Installing from source

If you are interested in contributing to Matplotlib development, running the latest source code, or just like to build everything yourself, it is not difficult to build Matplotlib from source. Grab the latest tar.gz release file from the PyPI files page, or if you want to develop Matplotlib or just need the latest bugfixed version, grab the latest git version Install from source.

The standard environment variables CC, CXX, PKG_CONFIG are respected. This means you can set them if your toolchain is prefixed. This may be used for cross compiling.

export CC=x86_64-pc-linux-gnu-gcc
export CXX=x86_64-pc-linux-gnu-g++
export PKG_CONFIG=x86_64-pc-linux-gnu-pkg-config

Once you have satisfied the requirements detailed below (mainly Python, NumPy, libpng and FreeType), you can build Matplotlib.

cd matplotlib
python -mpip install .

We provide a setup.cfg file which you can use to customize the build process. For example, which default backend to use, whether some of the optional libraries that Matplotlib ships with are installed, and so on. This file will be particularly useful to those packaging Matplotlib.

If you have installed prerequisites to nonstandard places and need to inform Matplotlib where they are, edit setupext.py and add the base dirs to the basedir dictionary entry for your sys.platform; e.g., if the header of some required library is in /some/path/include/someheader.h, put /some/path in the basedir list for your platform.

Dependencies

Matplotlib requires a large number of dependencies:

Optionally, you can also install a number of packages to enable better user interface toolkits. See What is a backend? for more details on the optional Matplotlib backends and the capabilities they provide.

  • tk (>= 8.3, != 8.6.0 or 8.6.1): for the Tk-based backends;
  • PyQt4 (>= 4.6) or PySide: for the Qt4-based backend;
  • PyQt5: for the Qt5-based backend;
  • pygtk (>= 2.4): for the GTK and the GTKAgg backend;
  • wxpython (>= 2.9 or later): for the WX or WXAgg backend;
  • cairocffi (>= v0.8): for cairo based backends;
  • pycairo: for GTK3Cairo;
  • Tornado: for the WebAgg backend;

For better support of animation output format and image file formats, LaTeX, etc., you can install the following:

Note

Matplotlib depends on a large number of non-Python libraries. pkg-config can be used to find required non-Python libraries and thus make the install go more smoothly if the libraries and headers are not in the expected locations.

Note

The following libraries are shipped with Matplotlib:

  • Agg: the Anti-Grain Geometry C++ rendering engine;
  • qhull: to compute Delaunay triangulation;
  • ttconv: a true type font utility.

Building on Linux

It is easiest to use your system package manager to install the dependencies.

If you are on Debian/Ubuntu, you can get all the dependencies required to build Matplotlib with:

sudo apt-get build-dep python-matplotlib

If you are on Fedora, you can get all the dependencies required to build Matplotlib with:

sudo dnf builddep python-matplotlib

If you are on RedHat, you can get all the dependencies required to build Matplotlib by first installing yum-builddep and then running:

su -c "yum-builddep python-matplotlib"

These commands do not build Matplotlib, but instead get and install the build dependencies, which will make building from source easier.

Building on macOS

The build situation on macOS is complicated by the various places one can get the libpng and FreeType requirements (MacPorts, Fink, /usr/X11R6), the different architectures (e.g., x86, ppc, universal), and the different macOS versions (e.g., 10.4 and 10.5). We recommend that you build the way we do for the macOS release: get the source from the tarball or the git repository and install the required dependencies through a third-party package manager. Two widely used package managers are Homebrew, and MacPorts. The following example illustrates how to install libpng and FreeType using brew:

brew install libpng freetype pkg-config

If you are using MacPorts, execute the following instead:

port install libpng freetype pkgconfig

After installing the above requirements, install Matplotlib from source by executing:

python -mpip install .

Note that your environment is somewhat important. Some conda users have found that, to run the tests, their PYTHONPATH must include /path/to/anaconda/.../site-packages and their DYLD_FALLBACK_LIBRARY_PATH must include /path/to/anaconda/lib.

Building on Windows

The Python shipped from https://www.python.org is compiled with Visual Studio 2008 for versions before 3.3, Visual Studio 2010 for 3.3 and 3.4, and Visual Studio 2015 for 3.5 and 3.6. Python extensions are recommended to be compiled with the same compiler.

Since there is no canonical Windows package manager, the methods for building FreeType, zlib, and libpng from source code are documented as a build script at matplotlib-winbuild.

There are a few possibilities to build Matplotlib on Windows:

Wheel builds using conda packages

This is a wheel build, but we use conda packages to get all the requirements. The binary requirements (png, FreeType,...) are statically linked and therefore not needed during the wheel install.

The commands below assume that you can compile a native Python lib for the Python version of your choice. See this howto for how to install and setup such environments. If in doubt: use Python >= 3.5 as it mostly works without fiddling with environment variables:

# create a new environment with the required packages
conda create  -n "matplotlib_build" python=3.5 numpy python-dateutil pyparsing pytz tornado "cycler>=0.10" tk libpng zlib freetype
activate matplotlib_build
# if you want a qt backend, you also have to install pyqt (be aware that pyqt doesn't mix well if
# you have created the environment with conda-forge already activated...)
conda install pyqt
# this package is only available in the conda-forge channel
conda install -c conda-forge msinttypes
# for Python 2.7
conda install -c conda-forge backports.functools_lru_cache

# copy the libs which have "wrong" names
set LIBRARY_LIB=%CONDA_PREFIX%\Library\lib
mkdir lib || cmd /c "exit /b 0"
copy %LIBRARY_LIB%\zlibstatic.lib lib\z.lib
copy %LIBRARY_LIB%\libpng_static.lib lib\png.lib

# Make the header files and the rest of the static libs available during the build
# CONDA_DEFAULT_ENV is a env variable which is set to the currently active environment path
set MPLBASEDIRLIST=%CONDA_PREFIX%\Library\;.

# build the wheel
python setup.py bdist_wheel

The build_alllocal.cmd script in the root folder automates these steps if you have already created and activated the conda environment.

Conda packages

This needs a working installed C compiler for the version of Python you are compiling the package for but you don't need to setup the environment variables:

# only the first time...
conda install conda-build

# the Python version you want a package for...
set CONDA_PY=3.5

# builds the package, using a clean build environment
conda build ci\conda_recipe

# install the new package
conda install --use-local matplotlib