svg.charts - Package for generating SVG Charts in Python

Status and License

svg.charts is a pure-python library for generating charts and graphs in SVG, originally based on the SVG::Graph Ruby package by Sean E. Russel.

svg.charts supercedes svg_charts 1.1 and 1.2.

svg.charts is written by Jason R. Coombs. It is licensed under an MIT-style permissive license.

You can install it with easy_install or pip:

easy_install svg.charts
pip install svg.charts

Or, check out the mercurial repository source.

Tests are continuously run by Travis-CI: BuildStatus

To run the tests, refer to the .travis.yml file for the steps run on the Travis-CI hosts.


svg.charts depends heavily on lxml and cssutils. Thanks to the contributors of those projects for stable, performant, standards-based packages.

Sean E. Russel for creating the SVG::Graph Ruby package from which this Python port was originally derived.

Leo Lapworth for creating the SVG::TT::Graph package which the Ruby port was based on.

Stephen Morgan for creating the TT template and SVG.

Getting Started

svg.charts has some examples (taken directly from the reference implementation) in tests/samples.py. These examples show sample usage of the various chart types. They should provide a good starting point for learning the usage of the library.

An example of using svg.charts in a CherryPy web app can be found in jaraco.site.charts. If the site is working, you can see the rendered output here.

svg.charts also provides API documentation.

Upgrade Notes

Upgrading from 1.x to 2.0

I suggest removing SVG 1.0 from the python installation. This involves removing the SVG directory (or svg_chart*) from site-packages.

Change import statements to import from the new namespace.

from SVG import Bar Bar.VerticalBar(...) becomes from svg.charts.bar import VerticalBar VerticalBar(...)

More To-Dos

  • Documentation! This package desperately needs some high-level, tutorial-style how-tos, and not just links to example code.
  • Implement javascript-based animation (See JellyGraph for a Silverlight example of what simple animation can do for a charting library).

Reporting Bugs and Getting Help

This project is hosted at sourceforge. Please use that site for reporting bugs and requesting help. Patches are also welcome.



  • #4: Added hook in Graph to allow overriding of the attributes on the root SVG element. One can now override or monkeypatch Graph._get_root_attributes to alter the rendering of the root attributes such as width and height. For example, to omit width and height:

    class MyPlot(plot.Plot):
        def _get_root_attributes(self):
            attrs = super(MyPlot, self)._get_root_attributes()
            del attrs['width']
            del attrs['height']
            return attrs


  • #1: Fixed javascript ID names in TimeSeries labels.


  • #5: Fixed references to class attributes in graph.py.


  • SF Issue #1: Fixed installation on Unix systems again. Author's preference for lowercase readme.txt was trumped by setuptools #100.
  • Moved hosting to BitBucket.
  • Established Continuous Integration Tests on Github mirror using Travis-CI.


  • Project now builds and tests pass on Python 3 without 2to3.


  • Corrected buggy logic in y-axis label rendering (thanks to Emmanuel Blot).
  • Converted to Unix line endings.


  • Updated to latest cssutils with Python 3 support. Thanks Christof!
  • Fixed a few remaining issues with Python 3 compatibility.


  • Fixed bug in rendering of Pie Chart styles.
  • Improved testing framework. Now samples are at least generated as part of the test suite.
  • Fixed bug in javascript when label ids had spaces. See #3139197.
  • Fixed build issue where package data wasn't included due to 2to3 technique. Now using distribute technique and installation on Python 3 requires distribute.


  • Fixed bug where x axis labels would not be rendered properly if the largest value was the same as the largest visible x value on the chart.


  • Altered the way CSS files are loaded, so they can be more easily customized by subclasses (and less dependent on the class names).


  • A small attempt to improve the documentation - added links to examples that already exist.


  • Fix IndexError in svg.charts.plot.Plot.field_size when there are only two values returned by float_range (in the case there are only two different 'y' values in the data) and scale_y_integers == True. Credit to Jean Schurger for the patch.
  • Fixed problem in setup.py installing on Unix OS (case sensitivity of readme.txt). Credit to Luke Miller and Jean Schurger for supplying a patch for this issue.


  • Updated cssutils dependency to 0.9.6 (currently in beta) to require the CSS profiles support.
  • Completed an SVG CSS profile according to the SVG 1.1 spec.


  • Added preliminary SVG CSS profile, suitable for stock CSS properties.


  • First major divergence from the Ruby reference implementation
  • Now implemented as a namespace package (svg.charts instead of svg_charts)
  • Changed XML processor to lxml
  • Enabled extensible css support using cssutils, greatly reducing static CSS
  • Renamed modules and methods to be more consistent with PEP-8 naming convention


  • Bug fixes


  • First public release