Download code from GitHub
A picture is worth a thousand words.
We all know that images are a powerful form of communication. We often use them to understand a situation better or to condense pieces of information into a graphical representation.
Just to give a couple of examples on how helpful they can be, let's consider the scientific and performance analysis fields. In order to clearly identify the bottlenecks, it is very important to be able to visualize data when analyzing performance information. Similarly, taking a quick glance at a graph drawn for a scientific experiment can give a scientist a better understanding of the results, something which is harder to achieve by looking only at the raw data.
Python is an interpreted language with a strong core functions basis and a powerful modular aspect which allows us to expand the language with external modules that offer new functionalities.
Modules reflect the Unix philosophy:
Do one thing, do it well.
So the result is that we have an extensible language with tools to accomplish a single task in the best possible way. Modules are often organized in packages. A package is a structured collection of modules that have the same purpose. One example of a package is Matplotlib.
Matplotlib is a Python package for 2D plotting that generates production-quality graphs. It supports interactive and non-interactive plotting, and can save images in several output formats (PNG, PS, and others). It can use multiple window toolkits (GTK+, wxWidgets, Qt, and so on) and it provides a wide variety of plot types (lines, bars, pie charts, histograms, and many more). In addition to this, it is highly customizable, flexible, and easy to use.
The dual nature of Matplotlib allows it to be used in both interactive and non-interactive scripts. It can be used in scripts without a graphical display, embedded in graphical applications, or on web pages. It can also be used interactively with the Python interpreter or IPython.
In this chapter, we will introduce Matplotlib, learn what it is, and what it can do. Later on, we will see what tools and Python modules are needed to have the best experience with Matplotlib and how to get them installed on our system, be it Linux, Windows, or Mac OS X.
The topics we are going to cover are:
Introduction to Matplotlib
Output formats and backends
Dependencies
How to install Matplotlib
The idea behind Matplotlib can be summed up in the following motto as quoted by John Hunter, the creator and project leader of Matplotlib:
Matplotlib tries to make easy things easy and hard things possible.
We can generate high quality, publication-ready graphs with minimal effort (sometimes we can achieve this with just one line of code or so), and for elaborate graphs, we have at hand a powerful library to support our needs.
Matplotlib was born in the scientific area of computing, where gnuplot and MATLAB were (and still are) used a lot.
With the entrance of Python into scientific toolboxes, an example of a workflow to process some data might be similar to this: "Write a Python script to parse data, then pass the data to a gnuplot script to plot it". Now with Matplotlib, we can write a single script to parse and plot data, with a lot more flexibility (that gnuplot doesn't have) and consistently using the same programming language.
We have to think of plotting not just as the final step in working with our data, but as an important way of getting visual feedback during the process. Here, the interactive capabilities of Matplotlib will come and rescue us.
Matplotlib was modeled on MATLAB, because graphing was something that MATLAB did very well. The high degree of compatibility between them made many people move from MATLAB to Matplotlib, as they felt like home while working with Matplotlib.
But what are the points that built the success of Matplotlib? Let's look at some of them:
It uses Python: Python is a very interesting language for scientific purposes (it's interpreted, high-level, easy to learn, easily extensible, and has a powerful standard library) and is now used by major institutions such as NASA, JPL, Google, DreamWorks, Disney, and many more.
It's open source, so no license to pay: This makes it very appealing for professors and students, who often have a low budget.
It's a real programming language: The MATLAB language (while being Turing-complete) lacks many of the features of a general-purpose language like Python.
It's much more complete: Python has a lot of external modules that will help us perform all the functions we need to. So it's the perfect tool to acquire data, elaborate the data, and then plot the data.
It's very customizable and extensible: Matplotlib can fit every use case because it has a lot of graph types, features, and configuration options.
It's integrated with LaTeX markup: This is really useful when writing scientific papers.
It's cross-platform and portable: Matplotlib can run on Linux, Windows, Mac OS X, and Sun Solaris (and Python can run on almost every architecture available).
In short, Python became very common in the scientific field, and this success is reflected even on this book, where we'll find some mathematical formulas. But don't be concerned about that, we will use nothing more complex than high school level equations.
The idea behind Matplotlib can be summed up in the following motto as quoted by John Hunter, the creator and project leader of Matplotlib:
Matplotlib tries to make easy things easy and hard things possible.
We can generate high quality, publication-ready graphs with minimal effort (sometimes we can achieve this with just one line of code or so), and for elaborate graphs, we have at hand a powerful library to support our needs.
Matplotlib was born in the scientific area of computing, where gnuplot and MATLAB were (and still are) used a lot.
With the entrance of Python into scientific toolboxes, an example of a workflow to process some data might be similar to this: "Write a Python script to parse data, then pass the data to a gnuplot script to plot it". Now with Matplotlib, we can write a single script to parse and plot data, with a lot more flexibility (that gnuplot doesn't have) and consistently using the same programming language.
We have to think of plotting not just as the final step in working with our data, but as an important way of getting visual feedback during the process. Here, the interactive capabilities of Matplotlib will come and rescue us.
Matplotlib was modeled on MATLAB, because graphing was something that MATLAB did very well. The high degree of compatibility between them made many people move from MATLAB to Matplotlib, as they felt like home while working with Matplotlib.
But what are the points that built the success of Matplotlib? Let's look at some of them:
It uses Python: Python is a very interesting language for scientific purposes (it's interpreted, high-level, easy to learn, easily extensible, and has a powerful standard library) and is now used by major institutions such as NASA, JPL, Google, DreamWorks, Disney, and many more.
It's open source, so no license to pay: This makes it very appealing for professors and students, who often have a low budget.
It's a real programming language: The MATLAB language (while being Turing-complete) lacks many of the features of a general-purpose language like Python.
It's much more complete: Python has a lot of external modules that will help us perform all the functions we need to. So it's the perfect tool to acquire data, elaborate the data, and then plot the data.
It's very customizable and extensible: Matplotlib can fit every use case because it has a lot of graph types, features, and configuration options.
It's integrated with LaTeX markup: This is really useful when writing scientific papers.
It's cross-platform and portable: Matplotlib can run on Linux, Windows, Mac OS X, and Sun Solaris (and Python can run on almost every architecture available).
In short, Python became very common in the scientific field, and this success is reflected even on this book, where we'll find some mathematical formulas. But don't be concerned about that, we will use nothing more complex than high school level equations.
The official Matplotlib presence on the Web is made up of two web sites:
The SourceForge project page at http://sourceforge.net/projects/matplotlib/
The main web site at http://matplotlib.sourceforge.net/
The SourceForge page contains, in particular, information about the development of Matplotlib, such as the released source code tarballs and binary packages, the SVN repository location, the bug tracking system, and so on. SourceForge also hosts some mailing lists for Matplotlib which are used for developers' discussions and users support.
On the main web site, we can find several important pieces of information about the Matplotlib package itself. For example:
The official documentation for Matplotlib is extensive. It covers in detail, all the submodules and the methods exposed by them, including all of their arguments. There are too many function arguments to cover in this book, so we are presenting only the most common ones here. In case of any doubts or questions, the official documentation is a good place to start your research or to look for an answer.
We encourage you to take a look at the gallery—it's inspiring!
The aim of Matplotlib is to generate graphs. So, we need a way to actually view these images or even to save them to files. We're going to look at the various output formats available in Matplotlib and the graphical user interfaces (GUIs) supported by the library.
Given its scientific roots (that means several different needs), Matplotlib has a lot of output formats available, which can be used for articles/books and other print publications, for web pages, or for any other reason we can think of. Let's first differentiate the output formats into two distinct categories:
Raster images: These are the classic images we can find on the Web or used for pictures. The most well known raster file formats are PNG, JPG, and BMP. They are widespread and well supported. The format of these images is like a matrix, with rows and columns, and at every matrix cell we have a pixel description (containing information such as colors). This format is said to be resolution-dependent, because the size of the matrix (the number of rows and columns) is determined when the image is created. An important parameter for raster images is the DPI(dots-per-inch) value. Once the image dimensions are decided (length and width, in inches), the DPI value specifies the detail level of the image. Hence, higher the DPI value, higher is the quality of the image (because for the same inch we get more dots). Scaling operations such as zooming or resizing can result in a loss of quality, because the image contains only a limited amount of information.
Vector images: As opposed to raster images, vector images contain a description of the image in the form of mathematical equations and geometrical primitives (for example, points, lines, curves, polygons, or shapes). We can think of this format as a series of directives to plot the image: "Draw a point here, draw another point there, draw a line between those two points" and so on. Given this descriptive format, these images are said to be resolution-independent, because it's the image interpreter that replots the image at the requested resolution using the instructions in it. Typical examples of vector image usage are typesetting and CAD (architectural or mechanical parts drawings).
Of course, Matplotlib supports both the categories, particularly with the following output formats:
|
Format |
Type |
Description |
|---|---|---|
|
EPS |
Vector |
Encapsulated PostScript. |
|
JPG |
Raster |
Graphic format with lossy compression method for photographic output. |
|
|
Vector |
Portable Document Format (PDF). |
|
PNG |
Raster |
Portable Network Graphics (PNG), a raster graphics format with a lossless compression method (more adaptable to line art than JPG). |
|
PS |
Vector |
Language widely used in publishing and as printers jobs format. |
|
SVG |
Vector |
Scalable Vector Graphics (SVG), XML based. |
PS or EPS formats are particularly useful for plots inclusion in LaTeX documents, the main scientific articles format since decades.
In the previous section, we saw the file output formats — they are also called hardcopy backends as they create something (a file on disk).
A backend that displays the image on screen is called a user interface backend.
The backend is that part of Matplotlib that works behind the scenes and allows the software to target several different output formats and GUI libraries (for screen visualization).
In order to be even more flexible, Matplotlib introduces the following two layers structured (only for GUI output):
The standard renderer is the Anti-Grain Geometry ( AGG) library, a high performance rendering engine which is able to create images of publication level quality, with anti-aliasing, and subpixel accuracy. AGG is responsible for the beautiful appearance of Matplotlib graphs.
The canvas is provided with the GUI libraries, and any of them can use the AGG rendering, along with the support for other rendering engines (for example, GTK+).
Let's have a look at the user interface toolkits and their available renderers:
|
Backend |
Description |
|---|---|
|
|
GTK+ (The GIMP ToolKit GUI library) canvas with AGG rendering. |
|
|
GTK+ canvas with GDK rendering. GDK rendering is rather primitive, and doesn't include anti-aliasing for the smoothing of lines. |
|
|
GTK+ canvas with Cairo rendering. |
|
|
wxWidgets (cross-platform GUI and tools library for GTK+, Windows, and Mac OS X. It uses native widgets for each operating system, so applications will have the look and feel that users expect on that operating system) canvas with AGG rendering. |
|
|
wxWidgets canvas with native wxWidgets rendering. |
|
|
Tk (graphical user interface for Tcl and many other dynamic languages) canvas with AGG rendering. |
|
|
Qt (cross-platform application framework for desktop and embedded development) canvas with AGG rendering (for Qt version 3 and earlier). |
|
|
Qt4 canvas with AGG rendering. |
|
|
FLTK (cross-platform C++ GUI toolkit for UNIX/Linux (X11), Microsoft Windows, and Mac OS X) canvas with Agg rendering. |
Here is the list of renderers for file output:
|
Renderer |
File type |
|---|---|
|
AGG |
|
|
PS |
|
|
|
|
|
SVG |
.svg |
|
Cairo |
|
|
GDK |
|
The renderers mentioned in the previous table can be used directly in Matplotlib, when we want only to save the resulting graph into a file (without any visualization of it), in any of the formats supported.
We have to pay attention when choosing which backend to use. For example, if we don't have a graphical environment available, then we have to use the AGG backend (or any other file). If we have installed only the GTK+ Python bindings, then we can't use the WX backend.
As mentioned earlier, Matplotlib has its origin in scientific fields, so it is commonly used to plot huge datasets. Python's native support for long lists becomes impractical for such sizes, so Matplotlib needs better support for arrays.
NumPy, the de facto standard Python module for numerical elaborations, provides support for high performance operations even with big mathematical data types such as arrays or matrices—along with many other mathematical functions that can be useful to Matplotlib users.
NumPy has to be available to use Matplotlib.
Once we have chosen the set of user interfaces (UIs) we prefer, then we need to install the Python bindings for them. Here is a summarizing list:
|
User Interface (UI) |
Binding |
Version |
Description |
|---|---|---|---|
|
FLTK |
pyFLTK |
1.0 or higher |
pyFLTK provides Python wrappers for the FLTK widgets library for use with |
|
GTK+ |
PyGTK |
2.2 or higher |
PyGTK provides Python wrappers for the GTK+ widgets library to use it with the It is recommended to use a version higher than 2.12, for a correct memory management. |
|
Qt |
PyQt or PyQt4 |
3.1 or higher and for Qt4, 4.0 or higher |
PyQt or PyQt4 provides Python wrappers for the Qt toolkit and is required by the Matplotlib |
|
Tk |
PyTK |
8.3 or higher |
Python wrapper for Tcl or Tk widgets library is used in |
|
Wx |
wxPython |
2.6 or higher, or 2.8 or higher |
wxPython provides Python wrappers for the wxWidgets library for use with the |
Another important tool, in particular for interactive usage, is IPython. It's an interactive Python shell with a lot of useful features, such as history, commands repeating, and others. It already has a Matplotlib mode in it. We'll be using IPython in this book, so it is recommended to install it.
Some of the tools that are needed by Matplotlib are already shipped with it (in the source code as well as in the binary distributions). Here is the list of those tools:
AGG (version 2.4): This is the Anti-Grain Geometry rendering engine. The local copy of the library is linked with the Matplotlib code in a static way. So, there's no need to install it (as a shared library).
pytz (version 2007g or higher): This is used for handling the time zone for
datetimePython objects. It will be installed if it's not already present in the system. It can be overridden usingsetup.cfg.python-dateutil (version1.1 or higher): This is used for enhanced handling of the
datetimePython objects. It needs to be installed if it's not already present in the system and can be overridden usingsetup.cfg.
The following tools are needed if we're going to install Matplotlib from the source:
Python: Currently, only Python 2.x is supported (no Python 3 yet)
NumPy: Version 1.1 or higher
libpng: Version 1.1 or higher is needed to load or save PNG images (Windows users can skip this requirement)
FreeType: Version 1.4 or higher is needed for reading TrueType font files (Windows users can skip this requirement)
There are several ways to install Matplotlib on our system:
Using packages from a Linux distribution
Using binary installers (for Windows and Mac OS X only)
Using packaged Python distributions that contain Matplotlib in the toolbox proposed
From the source code
We will look at each option in detail. We assume that Python, NumPy, and the optional build and runtime dependencies are already installed in the system (in order to install them, refer to their installation guides).
The advantage of using a Linux distribution is that several programs and libraries are already prepared by the distribution developers and made available (in a package format) to users. All we have to do is use the right tool and install the package.
In the following table, we will present some of the common Linux distributions package names for Matplotlib and the tools we can use to install the package:
|
Distribution |
Package name |
Installer tool |
|---|---|---|
|
Debian or Ubuntu (and all other Debian derivatives) |
|
Synaptic (graphical) apt-get or aptitude (command line) |
|
Fedora |
|
PackageKit (graphical) yum or rpm (command line) |
|
openSUSE |
|
YaST (graphical) zipper or rpm (command line) |
Before we can install Matplotlib, we have to satisfy its main dependencies. So, we have to download:
Installers for Python, which are available in the DOWNLOAD section of http://www.python.org/
Installers for NumPy, which are available in the Download section of http://www.scipy.org/
Once we've got the above packages correctly installed, we can go to the main project page of Matplotlib on SourceForge at http://sourceforge.net/projects/matplotlib/. In the Files section, we can find the relative versions of the binary packages for the Python that we have just installed (2.4, 2.5, or 2.6).
The procedure to install Matplotlib correctly on Mac OS X is similar to that of Windows.
First of all, we need to download:
Installers for Python, which are available in the DOWNLOAD section of http://www.python.org/ (the one already available in Mac OS X fives problem when using Matplotlib)
Installers for NumPy, which are available in the Download section of http://www.scipy.org/ or can be retrieved directly from http://pythonmac.org/
At this point, once they are correctly installed, we can download the binary installer from the download area of Matplotlib SourceForce page at http://sourceforge.net/projects/matplotlib/ or we can retrieve the version available at http://pythonmac.org/.
There are some packaged distributions of Python that contain Matplotlib in them, along with many other tools, such as IPython, NumPy, SciPy, and so on. These distributions will set up all the necessary things we need so that we can use Matplotlib on our machine. Some of the distributions are as follows:
Enthought Python Distribution (EPD): This package is available for Windows, Mac OS X, and Red Hat. We can download it from http://www.enthought.com/products/epd.php.
Python(x,y): This package is available for Windows and Ubuntu at http://www.pythonxy.com/.
Sage: This package is available for Linux at http://www.sagemath.org/.
These are mainly scientific distributions that install a lot of tools we don't directly need or use, but they have the advantage of making it easy to get Python, NumPy, and Matplotlib installed and working on our system.
There are two ways of obtaining the Matplotlib source code. They are:
Downloading it from the source code tarballs available in the download area of Matplotlib SourceForge project page at http://sourceforge.net/projects/matplotlib/.
Retrieving it from the Subversion (SVN) repository. This is the place where development takes place, so use it only if you know what you're doing.
If we decided to go with SVN, we can follow the instructions available in the Develop section of http://sourceforge.net/projects/matplotlib/.
If we are going to use the source code tarball, we will have to unpack it, go into the created source directory, and execute the following commands:
$ python setup.py build
$ sudo python setup.py install
These commands will build and then install Matplotlib. We will need administrative privileges to install it into the system directories (hence the sudo command in this Linux example).
Many aspects of the installation can be tuned using setup.cfg, a file shipped with the source code and used at build and install time. We can use it to customize the build process, such as changing the default backend, or choosing whether to install the optional libraries or not.
If we want to install Matplotlib from source on Windows, the Files section of Matplotlib SourceForge page contains handy egg files which we can download (choosing the Python version of interest) and then install using setuptools command. The following command will install Matplotlib on your machine:
$ easy_install matplotlib-<version>-py<py version>-win32.egg
Egg files are also available for Mac OS X, and we can use them in the same way as described above.
To ensure we have correctly installed Matplotlib and its dependencies, a very simple test can be carried out in the following manner:
$ python
Python 2.5.4 (r254:67916, Feb 18 2009, 03:00:47)
[GCC 4.3.3] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>> print numpy.__version__
1.2.1
>>> import matplotlib
>>> print matplotlib.__version__
0.98.5.3
If there's no error while executing this, then we are done.
In this chapter, we have covered the following areas:
What is Matplotlib and what are its main key points
The several file output formats and graphical user interfaces (GUIs) that are supported
The packages required by Matplotlib, and the ones needed for the GUI bindings
Installing and testing Matplotlib on a Linux, Windows, or Mac OS X system, in multiple ways
At this point, we only have a general idea of what Matplotlib is, along with the package correctly installed in our system. So let's go and start using Matplotlib!
