plot legends of a correlation matrix

Question

I have a correlation mat obtained from a data frame

>>> mat
                   Lcaud     Rcaud  Left_cereb_gm  Right_cereb_gm     Lamyg            
Rcaud           0.931934  1.000000       0.856891        0.715523  0.924995
Left_cereb_gm   0.915274  0.856891       1.000000        0.938301  0.601521
Right_cereb_gm  0.744007  0.715523       0.938301        1.000000  0.445450
Lamyg           0.754676  0.924995       0.601521        0.445450  1.000000
Rput            0.717757  0.876985       0.635881        0.462773  0.912815

I can nicely plot it with :

heatmap = plt.pcolor(mat, cmap=matplotlib.cm.Blues)

but I would like to get the legends 'Lcaud', 'Rcaud' etc... on the plot.

How can I achieve this?

Answer 1

I assume that you don't really want an actual legend since there is no easy way to define what color corresponds to each label or label-label pair. Instead, I think you are asking about setting the ticklabels? For example, given this data:

df = pd.DataFrame(np.random.rand(50, 5), columns=list('ABCDE'))
df.corr()
              A         B         C         D         E
A  1.000000  0.184661  0.125002  0.024962  0.115385
B  0.184661  1.000000  0.224653 -0.178857 -0.185907
C  0.125002  0.224653  1.000000 -0.011785  0.238073
D  0.024962 -0.178857 -0.011785  1.000000  0.171649
E  0.115385 -0.185907  0.238073  0.171649  1.000000

we can do the following:

import matplotlib.pyplot as plt
fig, ax = plt.subplots()
ax.pcolor(df.corr(), cmap=plt.cm.Blues)
ax.set_xticks(np.arange(5)+0.5) # center x ticks
ax.set_yticks(np.arange(5)+0.5) # center y ticks
ax.set_xticklabels(df.columns)
ax.set_yticklabels(df.columns)

which yields this:

Alternatively, Pandas has a helper function to help visualize the correlation among dataframe columns, called scatter_matrix:

from pandas import scatter_matrix
scatter_matrix(df, diagonal='kde')

yields