Use sharex=True in plotting seasonal decomposition #8514
Conversation
|
Is there an API call to disable it from an existing plot? If so, could you add a note that sharex is true by default. Axis synchronization can be disabled using ax.set_???? |
Good question, I hadn't thought of this. It took me a little digging to follow the layers of successive closed Matplotlib Issues and PR:
So the short answer seems to be "no, it's not possible to unshare axes", but perhaps I missed something cause I read the threads quite fast. Coming back to the display of decomposed time series, I don't think it's a problem though, for two reasons:
|
|
Do you want to add a keyword argument so that this is optional? I'm find with True being the default. |
|
Sorry for the late reply.
This is possible, but is it worth complexifying the API by an extra keyword? Of course it's only a small extra complexity, but on the other hand, I don't see a use case for having unsynced axes. More generally, there are many other small plot parameters that could be brought to the |
This is a tiny change in the
plotmethod ofDecomposeResult(in statsmodels.tsa.seasonal). I started to use it just a few minutes ago on a quite long time series (8760 samples), so that in a interactive Matlplotlib plot, it is useful to zoom horizontally. Unfortunately, the horizontal zooms are not in sync.Adding
sharex=Truein the call toplt.subplotsallows synchronizing the zoom between all seasonal components.An alternative would be to have this synchronization optional (one extra parameter to
DecomposeResult.plot), but I don't see a use case which needs dis-synchronization, what do you think?