Resize subplots using seaborn

Question

I have just recently started to use matplotlib and seaborn to plot my graphs. This is the code that I wrote so far

count = 1
l=[13,0,47,29,10]
plt.figure(figsize=(30,40))

for ww in l:
    temp_dict = defaultdict(list)
    entropies = list()
    for k,v in df.ix[ww].iteritems():
        e = 0
        for i in v:
            temp_dict[k].append(float(i))
            if not float(i) == 0:
                e += -1.0*float(i)*log10(float(i))
        entropies.append(e)

    y = entropies
    x=(range(len(entropies)))
    slope, intercept, r_value, p_value, std_err = stats.linregress(x,y)

    plt.subplot(len(lista_autori),2,count)
    tdf = pd.DataFrame.from_dict(temp_dict, orient='columns')
    a = tdf.dropna()
    sns.set(font_scale=2)
    #sns.factorplot(size=2, aspect=1)
    sns.heatmap(a,linewidths=.5,cmap="RdBu_r")
    plt.xlabel(ur'year')
    plt.ylabel(ur'$PACS$')
    count +=1

    plt.subplot(len(lista_autori),2,count)
    plt.plot(x,y)
    x1 = (range(28))
    y1 =  [slope*i + intercept for i in x1]
    plt.plot(x1,y1)
    count +=1


plt.tight_layout();

The result is the following:

I would like to resize each row, assigning 2/3 of the row to the left hand side picture, the remaining to the right one. I tried to look at the answer given here but I found some difficulties when I have to mix ax and seaborn's heatmap. Any help or other solutions?

Answer 1

The answer to the linked question directly tells you how to achieve a grid with unequal column width using GridSpec. I do not see much of a difference here, but maybe the following is more understandable for you, because it uses seaborn heatmaps, more than one row and no "ax".

import numpy as np
import matplotlib.pyplot as plt 
from matplotlib import gridspec
import seaborn as sns

# generate some data
x = np.arange(0, 10, 0.2)
y = np.sin(x)
X = np.random.rand(3,4)

fig = plt.figure(figsize=(8, 6)) 
gs = gridspec.GridSpec(5, 2, width_ratios=[2, 1]) 

for i in range(5):
    plt.subplot(gs[i*2+0])
    sns.heatmap(X)
    plt.subplot(gs[i*2+1])
    plt.plot(x,y)

plt.show()