CODEWITHSUNDEEP
pythonmatplotlibseabornvisualization
furbaw
furbaw

Reputation: 109

Seaborn Facet Grid 'col=' not creating plots side by side in Jupyter Notebook

I'm trying to plot two seaborn catplots side by side and not one on top of the other. In order to not share the x axis I had to create a FacetGrid instance and use the '.map_dataframe' function to produce two plots but it has produced them one above the other. (I was able to create the plots side by side just by using 'sns.catplot' however the 'sharex=False' argument didn't work and I ended up with all of the values across both x-axes.)

Here is my code:

grid = sns.FacetGrid(data = pcd_data,  col = 'Multi Storey Property')
grid.map_dataframe(sns.catplot, kind = 'bar', x="House Type ",
                   y="Percentage",  height = 6, hue="Last Balance Range Sort", 
                   palette = 'pastel', alpha = 0.6)
plt.close(1)
plt.show()

And it has produced the following:

Seaborn Facet Grid Rows not Columns

Any help would be greatly appreciated.

I'm using Python 3.8.3 Anaconda. Conda 4.9.1

Thanks

Upvotes: 1

Views: 1497

Answers (2)

StupidWolf
StupidWolf

Reputation: 46898

You can provide the columns to catplot, using an example dataset:

import seaborn as sns
data = sns.load_dataset("exercise")
data = data.groupby(["diet","kind","time"])["pulse"].agg("mean").reset_index()
data.head()

    diet    kind    time    pulse
0   no fat  rest    1 min   91.8
1   no fat  rest    15 min  92.2
2   no fat  rest    30 min  93.0
3   no fat  walking 1 min   95.6
4   no fat  walking 15 min  98.6

sns.catplot(x='kind',y='pulse',hue='time',kind='bar',col='diet',data=data)

enter image description here

If you have different x-axis, you can make the x label a string:

data = data[~((data.diet=="no fat") & (data.kind=="rest"))]
data['kind'] = data['kind'].astype('str')

And similar to @Alex solution, you call facet and map barplot onto it, the first argument is for x-axis, 2nd y-axis and 3rd hue:

g = sns.FacetGrid(data, col="diet",sharex=False,sharey=False)
g.map(sns.barplot, "kind", "pulse",'time',hue_order=data.time.unique())

enter image description here

Upvotes: 4

Alex
Alex

Reputation: 7045

You are using catplot which returns a FacetGrid. Instead you should use sns.barplot directly:

grid = sns.FacetGrid(data=pcd_data, col="Multi Storey Property")
grid.map_dataframe(
    sns.barplot,
    x="House Type ",
    y="Percentage",
    hue="Last Balance Range Sort",
    palette="pastel",
    alpha=0.6,
)
grid.set_axis_labels("House Type", "Percentage")
grid.add_legend()

Upvotes: 2

Related Questions