Summary
Pychart is a pure Python plotting package. It can generate most basic kinds of charts that you would expect, on a par with a typical spreadsheet application.
It's output is essentially PostScript; it can also directly generate PDF files. So, you have all the customary benefits of working with PostScript. If you have Ghostscript installed, Pychart can use it to generate PNG format.
The API is all script-oriented. That is, there is no WYSIWYG chart editor; simple scripts construct a chart out of Python data structures. Most parameters have sensible defaults, so it's easy to make simple charts. But most everything is parameterized so you can make elaborate charts if you want.
Pychart homepage has more information, and a passel of sample plots.
Commentary
After surveying scads of Python plotting packages I chose Pychart to analyze reams of test data for a project and it worked well for me.
Things I liked:
- The documentation is pretty good.
- No additional packages are needed. Every other plotting package I've seen that is this powerful requires installing other packages such as numPy or Python Imaging Library. Pychart works stand-alone, with the exception of conversion utilities if you want output other than postscript.
- Fine control over details of chart appearance. Simple charts are simple to produce, and complicated charts are possible.
Things I disliked:
- There is some learning curve if you want to go beyond simple plots.
- The architecture is "postscript-flavored" (although you don't need to know any postscript to use Pychart, if you can think in Postscript, the API makes more sense.)
- The API has some design quirks. It's split into a lot of modules, most of which hold a single class 'T'. For example, Pychart.bar_plot.T, Pychart.line_plot.T, ....
[DanEverhart 2004/10/23]
(add more comments here)