Missing data results in misaligned plots

📊 Plotly Python
1

I’m attempting to plot several years of average temperatures (by month) in a line graph like so:

2021-02-22_7-45-13
2021-02-22_7-45-13852×401 25.3 KB

in the screenshot, you can notice that the 2012 line is a similar shape to the average, but seems to be missing 11 & 12 (November & December).

Actually, the data is missing January & February like so:

year month TMR_SUB_18
2012 3 23.61949528936744
2012 4 39.28492329149238
2012 5 49.55534946236562
2012 6 60.311488178025115
2012 7 60.60436827956987
2012 8 60.590900537634425
2012 9 50.8024712643678
2012 10 37.558411843876144
2012 11 22.451902777777793
2012 12 10.810748317631226

there are several periods of missing data in the dataset (which is growing with time) so manually filling in “filler” data is not a preferred option.

ideally, i’d prefer the month of data to match up with the months on the xaxis:

This is how i produce the original graph

fig = go.Figure()

for year in data_test_group.year.unique():
    # print(data_test_group.loc[data_test_group.year==year].groupby(['year', 'month'])['TMR_SUB_18'].mean())
    fig.add_trace(go.Scatter(x=month_list, 
                             y=data_test_group.loc[data_test_group.year==year].groupby(['year', 
                                                                                        'month'])['TMR_SUB_18'].mean(),
                    mode='lines+markers',
                    name=str(year) + ' Monthly Averages',
                            connectgaps=True,
                            line_shape='spline'))
    
    fig.update_xaxes(nticks=12)
    fig.update_xaxes(tick0=1, dtick=1)
    
fig.add_trace(go.Scatter(x=month_list, 
                         y=data_test_group.groupby(['month'])['TMR_SUB_18'].mean(), 
                         line = dict(color='firebrick', 
                                     width=4, 
                                     dash='dot'),
                        name='Average'))
fig.show()

please advise! Thank you in advance!

2

The month_list you’re using as the x axis is probably a string. I’m having the same problem and hope plotly fixes this sometime soon, but until then you can solve this by casting these values to datetime using pd.to_datetime. It seems that plotly respects the date order.