How to display a categorical dataframe in a pairplot

Question

I am getting the following error: No variables found for grid columns? In the console it prints the dataframe ok, but when I want to display it in seaborn, it doesn't work.

%matplotlib inline
import seaborn as sns
from ipywidgets import interact
import matplotlib.pyplot as plt

cars_url = 'http://archive.ics.uci.edu/ml/machine-learning-databases/car/car.data'
cars_columns = ['buying', 'maint', 'doors', 'persons', 'lug_boot', 'safety', 'class']
data = pd.read_csv(cars_url, names = cars_columns)
data1 = pd.DataFrame(data)
plot(data1)

@interact(hue = ['buying'])
def plot(hue):
    _ = sns.pairplot(data1, hue = hue)

`data1 = pd.DataFrame(data)` is not necessary. `data = pd.read_csv(cars_url, names = cars_columns)` is already a `DataFrame` — Trenton McKinney, Nov 11 '22 at 16:06

Andrew · Accepted Answer · 2022-11-11T23:54:28.460

3

Pairtplot works only with numerical data. You should use Label Encoder before paiplot.

For example:

from sklearn.preprocessing import OrdinalEncoder

encoder = OrdinalEncoder()
data = encoder.fit_transform(data)

edited Nov 11 '22 at 23:54

answered Nov 11 '22 at 14:55

Andrew

176
2
8

Thanks. It somehow helped me. I will convert everything to numerical data. – Paul Viorel Nov 11 '22 at 15:14
This does not work as written. It results in `ValueError: y should be a 1d array, got an array of shape (1728, 7) instead.` – Trenton McKinney Nov 11 '22 at 16:05
@PaulViorel if this is used on each column, the transformed values will be reused in each column. `df.buying.unique() → array([3, 0, 2, 1])` and `df.maint.unique() → array([3, 0, 2, 1])`. – Trenton McKinney Nov 11 '22 at 16:28
Y, you're right. Fixed that. – Andrew Nov 11 '22 at 23:54

How to display a categorical dataframe in a pairplot

1 Answers1