CODEWITHSUNDEEP
pythonjupyterseaborn
Sledge
Sledge

Reputation: 1345

Relabel axis ticks in seaborn heatmap

I have a seaborn heatmap that I am building from a matrix of values. Each element of the matrix corresponds to an entitiy that I would like to make the tick label for each row/col in the matrix.

I tried using the ax.set_xticklabel() function to accomplish this but it seems to do nothing. Here is my code:

type(jr_matrix)
>>> numpy.ndarray

jr_matrix.shape
>>> (15, 15)

short_cols = ['label1','label2',...,'label15'] # list of strings with len 15

fig, ax = plt.subplots(figsize=(13,10)) 
ax.set_xticklabels(tuple(short_cols)) # i also tried passing a list
ax.set_yticklabels(tuple(short_cols))
sns.heatmap(jr_matrix, 
            center=0, 
            cmap="vlag", 
            linewidths=.75, 
            ax=ax,
            norm=LogNorm(vmin=jr_matrix.min(), vmax=jr_matrix.max()))

The still has the matrix indices as labels:

enter image description here

Any ideas on how to correctly change these labels would be much appreciated.

Edit: I am doing this using jupyter notebooks if that matters.

Upvotes: 1

Views: 6862

Answers (1)

DavidG
DavidG

Reputation: 25371

You are setting the x and y tick labels of the axis you have just created. You are then plotting the seaborn heatmap which will overwrite the tick labels you have just set.

The solution is to create the heatmap first, then set the tick labels:

fig, ax = plt.subplots(figsize=(13,10)) 

sns.heatmap(jr_matrix, 
            center=0, 
            cmap="vlag", 
            linewidths=.75, 
            ax=ax,
            norm=LogNorm(vmin=jr_matrix.min(), vmax=jr_matrix.max()))

# passing a list is fine, no need to convert to tuples
ax.set_xticklabels(short_cols)
ax.set_yticklabels(short_cols)

Upvotes: 6

Related Questions