fill_between in line_plot still sees spike behavior

[gidden]

df here is queried of all iamc15 data as so

df = pyam.read_iiasa_iamc15(
    model='*', scenario='*',
    variable=['Emissions|CO2', 'Primary Energy|Coal', 
              'AR5 climate diagnostics|Temperature|Global Mean|MAGICC6|MED'], 
    region='World'
)

[danielhuppmann]

We had the same issue in the scenario explorer development. The solution there was to get the list of all years with at least one data point and then interpolate all timeseries for all these years.