如何在 seaborn 中的 lmplot 中添加比较线？

Question

I would like to combine the following lmplots I have. More specifically, the red lines are averages for each season, and I want to place them on their respective lmplots with the other data, instead of having them separate. Here is my code (note, the axes limits aren't working because the second lmplot is messing it up. It works when I just plot the initial data):

ax = sns.lmplot(data=data, x='air_yards', y='cpoe',col='season', lowess = True, scatter_kws={'alpha':.6, 'color': '#4F2E84'}, line_kws={'alpha':.6, 'color': '#4F2E84'})

ax = sns.lmplot(data=avg, x='air_yards', y= 'cpoe',lowess=True, scatter=False, line_kws={'linestyle':'--', 'color': 'red'}, col = 'season')

axes.set_xlim([-5,30])
axes.set_ylim([-25,25])

ax.set(xlabel='air yards')

And here is the output. Simply put, I want to take those red lines and put them on their respective year plots above. Thanks!

Answer 1

Not sure if it is possible the way you want, so maybe something like:

import matplotlib.pyplot as plt
import seaborn as sns

#dummy example
data = pd.DataFrame({'air_yards': range(1,11), 
                     'cpoe': range(1,11), 
                     'season': [1,2,3,2,1,3,2,1,3,2]})
avg = pd.DataFrame({'air_yards': [1, 10]*3, 
                    'cpoe': [2,2,5,5,8,8], 
                    'season': [1,1,2,2,3,3]})

# need this info
n = data["season"].nunique()

# create the number of subplots
fig, axes = plt.subplots(ncols=n, sharex=True, sharey=True)

# now you need to loop through unique season
for ax, (season, dfg) in zip(axes.flat, data.groupby("season")):
    # set title
    ax.set_title(f'season={season}')

    # create the replot for data
    sns.regplot("air_yards", "cpoe", data=dfg, ax=ax, 
                lowess = True, scatter_kws={'alpha':.6, 'color': '#4F2E84'}, 
                line_kws={'alpha':.6, 'color': '#4F2E84'})

    # create regplot for avg
    sns.regplot("air_yards", "cpoe", data=avg[avg['season'].eq(season)], ax=ax, 
                lowess=True, scatter=False, 
                line_kws={'linestyle':'--', 'color': 'red'})

plt.show()

you get