Since I will not know in advance the values users are

> plotting, I cannot directly set the plot limits as far as

> I know. Any advice on how best to handle this problem? I

> could try to calculate a best limit for each plot, or

> simply use a single large value such that the Y axis label

> is never pushed off the graph. Any other advice or ideas?

This is basically the best approach. You can make the default

subplots_adjust(left=.1) or something like that to give yourself

extra room. matplotlib will fall over to exponential ticking above

10^4, so 1000 is pretty much your worst case. Just make sure you can

fit the ylabel on this case and you should be OK

pylab.scatter([1000, 5000], [1000, 5000])

pylab.subplots_adjust(left=0.1)

Let us know when the new plots are up!

I assume you are using

import matplotlib

matplotlib.use('Agg')

for the main site and not GTKAgg....

JDH