Graphic file formats

[Rich Shepard]

On Thu, 9 Dec 2021, Neal Becker via lyx-users wrote:

> I've had some issues similar to this, although none of my timings are as
> long as you describe. I often create plots with matplotlib. I believe that
> if you draw, say a scatterplot with 10^6 data points, the resulting pdf
> will contain 10^6 instructions to draw points. This may cause slow
> rendering (and printing!). The most straightforward fix I know of is don't
> do that, write code to resample the results to a smaller set. However,
> that requires work on my part.

Neal,

The data used to create a plot should be appropriate to communicate what's
important about the data. Recently I calculated monthly averages for river
discharge and plotted those because the data set was very large and monthly
averages were sufficient for that project.

My current project focuses on a different river's discharge variability at
multiple time scales (hourly, daily, monthly, seasonally, and annually) so I
need all 813694 measurements over the 33 year period of record for an
overall data distribution summary.

> I've searched a couple of times for a software solution (say, a backend
> for mpl that would automatically simplify plots), but so far I haven't
> found anything. Similarly, it would be nice to have a solution that
> directly simplifies a pdf. I don't think this is the same thing as
> compressing a pdf (although I could be wrong).

I've no idea what 'simplifying' a plot would entail. But, since ghostscript
reduced my plot's size by about 74% with no effect on resolution I'm
satisfied with that.

Regards,

Rich