axvline plot is not visible when one signal is plotted - matplotlib python

Question

I am trying to plot certain discrete scores based on a timestamp at which they were recorded. There are 4 separate dataframes(pandas) called lower-back-Now, lower-back-Worst, Neck-Now, and Neck-Worst. And also, at some particular days offset from the initial timestamp index I plot axvlines to denote some interventions with text. The dataframes look like these..

lbNow
                    Answer
StartUnixTime             
1970-01-01 16:45:55      4
1970-01-01 19:30:53      5
1970-01-02 05:00:11      6
1970-01-02 19:31:01      3
1970-01-03 05:06:29      5

lbWst
                    Answer
StartUnixTime             
1970-01-01 16:46:08      6
1970-01-01 19:30:53      5
1970-01-02 05:00:18      6
1970-01-02 19:31:10      7
1970-01-03 05:06:29      5

neckNow
                    Answer
StartUnixTime             
1970-01-01 16:46:26      4
1970-01-01 19:31:39      6
1970-01-02 05:00:27      6
1970-01-02 19:31:17      2
1970-01-03 05:07:04      4

neckWst       
                    Answer
StartUnixTime             
1970-01-01 16:46:31      5
1970-01-01 19:31:47      6
1970-01-02 05:00:28      6
1970-01-02 19:31:22      7
1970-01-03 05:07:08      4

and the code for the plot for plottin using matplotlib in subplots is as follows:

fig = plt.figure(figsize=(15,15))
ax1 = fig.add_subplot(211)

myfmt = mdates.DateFormatter('%m-%d')

ax1.plot(lbNow,color='#055F65',label='Lower Back Now')
ax1.plot(lbWst,color='#c50d63',label='Lower Back worst',linestyle='--')
ax1.xaxis.set_major_formatter(myfmt)
for label in ax1.xaxis.get_ticklabels():
    label.set_rotation(45)
# fig.autofmt_xdate()

ax1.legend(loc='best')
ax1.set(title='Lower Back Pain Dairy',xlabel='Timestamps since start date',ylabel='Pain Scores')

ax1.axvline(x=lbNow.index[0] + pd.DateOffset(days=1),color = 'r',linestyle='-.')
ax1.text(lbNow.index[0] + pd.DateOffset(days=2),8.1,'MBB#1',rotation=45)

ax1.axvline(x=lbNow.index[0] + pd.DateOffset(days=22),color = 'b',linestyle='-.')
ax1.text(lbNow.index[0] + pd.DateOffset(days=23),8.1,'MBB#2',rotation=45)

ax1.axvline(x=lbNow.index[0] + pd.DateOffset(days=43),color = 'k',linestyle='-.')
ax1.text(lbNow.index[0] + pd.DateOffset(days=44),8.1,'RFA',rotation=45)
# ax1.set_ylim((0,10))
ax1.grid(True)

ax3 = fig.add_subplot(212)

ax3.plot(neckNow,color='#4b0d2b',label='Neck Now',linestyle='-')
ax3.plot(neckWst,color='#c50d63',label='Neck Worst',linestyle='--')

ax3.xaxis.set_major_formatter(myfmt)
for label in ax3.xaxis.get_ticklabels():
    label.set_rotation(45)
fig.autofmt_xdate()

ax3.legend(loc='lower center')
ax3.set(title='Neck Pain Dairy',xlabel='Timestamps since start date', ylabel='Pain Scores')

ax3.axvline(x=neckNow.index[0] + pd.DateOffset(days=1),color = 'r',linestyle='-.',clip_on=False)
ax3.text(neckNow.index[0] + pd.DateOffset(days=2),8,'MBB#1',rotation=45)

ax3.axvline(x=neckNow.index[0] + pd.DateOffset(days=22),color = 'b',linestyle='-.')
ax3.text(neckNow.index[0] + pd.DateOffset(days=23),8,'MBB#2',rotation=45)

ax3.axvline(x=neckNow.index[0] + pd.DateOffset(days=43),color = 'k',linestyle='-.')
ax3.text(neckNow.index[0] + pd.DateOffset(days=44),8,'RFA',rotation=45)

ax3.grid(True)
fig.tight_layout()
plt.show()

The problem is with the second subplot with neckNow and neckWst plotting. If I plot both df then the ax3.axvline do not appear as shown

But then if Comment the neckNow plot then the lines appear as shown.

Is there any problem with the y limits or is the timestamp causing this weird error ?

You can look into [z-order](https://matplotlib.org/2.0.2/examples/pylab_examples/zorder_demo.html) — harvpan, Jun 13 '18 at 16:09
Hey thanks for the suggestion, but no luck. The first subplot is working without z-order, so i dont think zorder is the issue. Nevertheless i did give a try but not change — lamo_738, Jun 13 '18 at 16:38
Which version of matplotlib (and other relevant libraries) are you using? I've tested your code with matplotlib v.2.2.2 and pandas v.0.23.1 all seems to work. — PieCot, Jun 13 '18 at 18:57
I am guessing the clip_on=False found in the line `ax3.axvline(x=neckNow.index[0] + pd.DateOffset(days=1),color = 'r',linestyle='-.',clip_on=False)` ,should be added to all the `ax.axvline` plotting lines. Or try removing them from all the `ax.axvline` plotting line....Take a look at [https://stackoverflow.com/questions/6146290/plotting-a-line-over-several-graphs?utm_medium=organic&utm_source=google_rich_qa&utm_campaign=google_rich_qa] for a similar use-case — neutralCreep, Jun 13 '18 at 19:07
@PieCot I am using matplotlib v2.1.2 and pandas v0.22.0 . I am using it under anaconda distribution. Should I need to update those ? .. so the code is working fine right ? — lamo_738, Jun 13 '18 at 19:54
@neutralCreep Hey, I did go through that post, and I did play around with clip_on param but there was no change. And also, in my problem statement, i dont want it across the two plots as done in the link you shared. — lamo_738, Jun 13 '18 at 19:55
@lamo_738 I'm running your code in a conda virtual environment too. If updating your libraries does not solve the problem, maybe you should check your data: it could contain some weird or unexpected value. — PieCot, Jun 13 '18 at 21:04
Hey .. I did update the lib, but there was no luck. But i did find out why ! apparently the 'Answer' pandas column was in string rather than integer. Hence, the y axis was a string rather than int between 0-10. So the y axis was screwing up the plots and axvline. Anyways Thanks for the help Guys ! — lamo_738, Jun 15 '18 at 17:29

axvline plot is not visible when one signal is plotted - matplotlib python

0 Answers0