conditional matplotlib fill_between for dataframe - pandas

Question

It seemed easy, but I struggle.
I want to color between axes conditionally, i.e. to have the area filled either green or red and fully filled. I use some boolean df column to determine colour
df = pd.DataFrame([1,1,1,1,1,0,0,1,1,0,1,1,0,1,0,1], columns=['grow'])
df['green']=(df.grow>0) | (df.grow.shift(1)>0)
df['red']= (df.grow<=0) | (df.grow.shift(1)<=0)
But when I use this condition to fill between with this:
fig, axes = plt.subplots(nrows=1, ncols=1)
axes.fill_between(df.index, 0, 1,
where=(df.grow>0) , color = 'green', alpha = 0.1)
axes.fill_between(df.index, 0, 1,
where=(df.grow<=0) , color = 'red', alpha = 0.1)
they are not fully filled. How shall I transform the limit in where to get good filling?

Answer 1

The reason why your filled areas are non-continuous becomes apparent when you plot pd.grow itself as a line plot. With the way you are using fill_between(), you are implicitly assuming that your data resembles a step function, but it is actually more of a saw tooth (the edges are not 'sharp'). One way to get around this is to fill the function with more, repeating values and thus make the transitions between 0s and 1s sharper. numpy is a practical tool for these kind of operations. Here an example of how to do it:
from matplotlib import pyplot as plt
import pandas as pd
import numpy as np
df = pd.DataFrame([1,1,1,1,1,0,0,1,1,0,1,1,0,1,0,1], columns=['grow'])
df['green']=(df.grow>0) | (df.grow.shift(1)>0)
df['red']= (df.grow<=0) | (df.grow.shift(1)<=0)
fig, axes = plt.subplots(
nrows=2, ncols=2, gridspec_kw = {'height_ratios':[1, 3]}
)
axes[0,0].plot(df.index, df.grow)
axes[0,0].set_title('original function')
axes[1,0].fill_between(df.index, 0, 1,
where=(df.grow>0) , color = 'green', alpha = 0.1)
axes[1,0].fill_between(df.index, 0, 1,
where=(df.grow<=0) , color = 'red', alpha = 0.1)
axes[1,0].set_title('original shading')
N=100
x = np.linspace(df.index[0],df.index[-1],N*len(df.index))
y = np.repeat(df.grow, N)
axes[0,1].plot(x,y)
axes[0,1].set_title('sharper step function')
axes[1,1].fill_between(x, 0, 1,
where=(y>0) , color = 'green', alpha = 0.1, lw=0)
axes[1,1].fill_between(x, 0, 1,
where=(y<=0) , color = 'red', alpha = 0.1,lw=0)
axes[1,1].set_title('new_shading')
plt.show()
...and the result looks like this:
Hope this helps.