Dear Nico,
Instead of "sonifying the plot", it might be interesting to generate both the plot and the sound from the same data?
How many samples are in your original time series? Could you play the series directly as an audio file (or use it to modulate a synthesis algorithm)?