Different y scale for each row Matplotlib

Question

I'm trying to dynamically scale the y-axes for each row of subplots, so i figured I would manually define the ymax for each subplot I would create with the for loop.

fig, ax = plt.subplots(3, len(motif.strip()), figsize=(15, 10), sharex=True, sharey=True)

    for i, s in enumerate(dataw.pos.unique()):
        for j, r in enumerate(dataw.type.sort_values().unique()):
            atmp = dataw[(dataw.pos == s) & (dataw.type == r)]
            btmp = dataw[(dataw.type == r)]
            ymax = (btmp['values'].values).max()
            #print(btmp)
            #print(ymax)
            tmp = [atmp[atmp['base'] == 'A']['values'].values,
                   atmp[atmp['base'] == 'G']['values'].values,
                   atmp[atmp['base'] == 'T']['values'].values,
                   atmp[atmp['base'] == 'C']['values'].values]

            ax[j][i].violinplot(tmp)
            ax[j][i].set_ylim([0, ymax])

            #ax[j][i].set(xlabel='base',
            #             ylabel='values',
            #             title=s + '--' + r)

    fig.tight_layout()

    fig.savefig(str(graph) + ".png")

What I have right now:

What I would like to have (note different y axes for each row):

Answer 1

Setting ylim manually won't work in your case sense sharey=True . If you set it to False, you can then set ylim for each plot (with separate y-ticks for each plot).

Since it seems you want a shared y for each row , you will have to calculate the max value for each row and set it outside of the internal loop. Then, you can hide the ticks for all but the first plot in each row: ax.set_yticks([]) .

Update

I don't know if this feature existed at the time I wrote the original answer and I was just not aware of it, but the correct way of achieving this today is to use: sharey='row' when calling the subplots function. See here for more details.