Bugzilla – Bug 14657
stars plot provides no way to distinguish lines
Last modified: 2011-08-16 14:35:23 UTC
I was trying to do a spider or radar plot of a few sets of data (including directly from the examples) and found that I could not visually distinguish any of the fields on the chart. A user on SO provided a simple patch that allowed me to color the lines.
Details and suggestion here:
I'll restrain myself and won't go into why I think you're using the wrong plot, but I will do ask you to, please, supply a proper patch against R including the documentation in the report. Please see the FAQ about how to report bugs and enhancements. Thanks.
(In reply to comment #1)
> I'll restrain myself and won't go into why I think you're using the wrong plot,
> but I will do ask you to, please, supply a proper patch against R including the
> documentation in the report. Please see the FAQ about how to report bugs and
> enhancements. Thanks.
Hmm.. about your restraint :
These kind of plots have lately become very popular, e.g. in Swiss "politics metrics" in order to characterize politicians, political parties, and even (potential) voters.
There, its a good idea to measure more than just one or two coordinates in the space of political views.
The guys who started propagating it, came from politology/geography and called it "smartspider" and now provide a very successful website "smartvote.ch".
What better graphics (for newspaper, web graphic, and the general not-well versed public) would you choose instead?
Martin, I don't think what you have in mind (or I hope at least) is Dustin's plot - what he wants is a essentially a less legible polar PCP. I'll be happy to discuss this but I don' think a bug reporting system is the right venue.
Created attachment 1223 [details]
Added ability to color lines in stars.
I couldn't find anything that looked like concise "how to contribute" documentation. Apologies if this is not acceptable.
While I'm certainly interested in finding better ways to display my data (the consumers of this data specifically requested this), I don't think there's an argument that if this type of plot is supported, having ways to distinguish the individual plotted items is important.
This patch achieves this along with documentation and an example.
Thank you for the patch. The format was fine.
As the change is really minimal,
(and in spite of or in light of (:-) Brian's remark in the useR! talk today),
I've committed it (+ a NEWS entry) for R-devel.