#import libraries

import seaborn as sns

import pandas as pd

# creating df
df=sns.load_dataset('tips')

df.head(5)

total_bill tip sex smoker day time size

0 16.99 1.01 Female No Sun Dinner 2
1 10.34 1.66 Male No Sun Dinner 3
2 21.01 3.50 Male No Sun Dinner 3
3 23.68 3.31 Male No Sun Dinner 2
4 24.59 3.61 Female No Sun Dinner 4

df.dtypes

total_bill float64
tip float64
sex category
smoker category
day category
time category
size int64
dtype: object

df.isnull().sum()

total_bill 0
tip 0
sex 0
smoker 0
day 0
time 0
size 0
dtype: int64

df.shape

(244, 7)

##Extract categorical columns from the dataframe

#Here we extract the columns with category datatype as they are the
categorical columns
categorical_columns =
df.select_dtypes(include=['category']).columns.tolist()

categorical_columns

['sex', 'smoker', 'day', 'time']

# ccreating features
X=df.drop('tip',axis=1)

total_bill sex smoker day time size

0 16.99 Female No Sun Dinner 2
1 10.34 Male No Sun Dinner 3
2 21.01 Male No Sun Dinner 3
3 23.68 Male No Sun Dinner 2
4 24.59 Female No Sun Dinner 4
.. ... ... ... ... ... ...
239 29.03 Male No Sat Dinner 3
240 27.18 Female Yes Sat Dinner 2
241 22.67 Male Yes Sat Dinner 2
242 17.82 Male No Sat Dinner 2
243 18.78 Female No Thur Dinner 2

[244 rows x 6 columns]

# creating target varibale

y=df['tip']

0 1.01
1 1.66
2 3.50
3 3.31
4 3.61
...
239 5.92
240 2.00
241 2.00
242 1.75
243 3.00
Name: tip, Length: 244, dtype: float64

#one hot encoding using OneHotEncoder of Scikit-Learn

from sklearn.preprocessing import OneHotEncoder

#Initialize OneHotEncoder
encoder=OneHotEncoder(sparse=False,dtype=int)

encoder

OneHotEncoder(dtype=<class 'int'>, sparse=False)

# Apply one-hot encoding to the categorical columns

one_hot_encoded=encoder.fit_transform(df[categorical_columns])#--
>df[['sex','day']]
C:\Users\lucky\anaconda3\Lib\site-packages\sklearn\preprocessing\
_encoders.py:868: FutureWarning: `sparse` was renamed to
`sparse_output` in version 1.2 and will be removed in 1.4.
`sparse_output` is ignored unless you leave `sparse` to its default
value.
warnings.warn(

# df[['sex','day']]
#df[categorical_columns]

#Create a DataFrame with the one-hot encoded columns

#We use get_feature_names_out() to get the column names for the
encoded data
encoded_df=pd.DataFrame(one_hot_encoded,columns=encoder.get_feature_na
mes_out(categorical_columns))

encoded_df

sex_Female sex_Male smoker_No smoker_Yes day_Fri day_Sat

day_Sun \
0 1 0 1 0 0 0
1
1 0 1 1 0 0 0
1
2 0 1 1 0 0 0
1
3 0 1 1 0 0 0
1
4 1 0 1 0 0 0
1
.. ... ... ... ... ... ...
...
239 0 1 1 0 0 1
0
240 1 0 0 1 0 1
0
241 0 1 0 1 0 1
0
242 0 1 1 0 0 1
0
243 1 0 1 0 0 0
0

day_Thur time_Dinner time_Lunch

0 0 1 0
1 0 1 0
2 0 1 0
3 0 1 0
4 0 1 0
.. ... ... ...
239 0 1 0
240 0 1 0
241 0 1 0
242 0 1 0
243 1 1 0

[244 rows x 10 columns]

# Concatenate the one-hot encoded dataframe with the original

dataframe
df_encoded = pd.concat([df, encoded_df], axis=1)

df_encoded

total_bill tip sex smoker day time size

sex_Female \
0 16.99 1.01 Female No Sun Dinner 2 1

1 10.34 1.66 Male No Sun Dinner 3 0

2 21.01 3.50 Male No Sun Dinner 3 0

3 23.68 3.31 Male No Sun Dinner 2 0

4 24.59 3.61 Female No Sun Dinner 4 1

.. ... ... ... ... ... ... ... ...

239 29.03 5.92 Male No Sat Dinner 3 0

240 27.18 2.00 Female Yes Sat Dinner 2 1

241 22.67 2.00 Male Yes Sat Dinner 2 0

242 17.82 1.75 Male No Sat Dinner 2 0

243 18.78 3.00 Female No Thur Dinner 2 1

sex_Male smoker_No smoker_Yes day_Fri day_Sat day_Sun

day_Thur \
0 0 1 0 0 0 1
0
1 1 1 0 0 0 1
0
2 1 1 0 0 0 1
0
3 1 1 0 0 0 1
0
4 0 1 0 0 0 1
0
.. ... ... ... ... ... ...
...
239 1 1 0 0 1 0
0
240 0 0 1 0 1 0
0
241 1 0 1 0 1 0
0
242 1 1 0 0 1 0
0
243 0 1 0 0 0 0
1

time_Dinner time_Lunch
0 1 0
1 1 0
2 1 0
3 1 0
4 1 0
.. ... ...
239 1 0
240 1 0
241 1 0
242 1 0
243 1 0

[244 rows x 17 columns]

# Drop the original categorical columns

df_encoded = df_encoded.drop(categorical_columns, axis=1)

df_encoded

total_bill tip size sex_Female sex_Male smoker_No

smoker_Yes \
0 16.99 1.01 2 1 0 1
0
1 10.34 1.66 3 0 1 1
0
2 21.01 3.50 3 0 1 1
0
3 23.68 3.31 2 0 1 1
0
4 24.59 3.61 4 1 0 1
0
.. ... ... ... ... ... ... .
..
239 29.03 5.92 3 0 1 1
0
240 27.18 2.00 2 1 0 0
1
241 22.67 2.00 2 0 1 0
1
242 17.82 1.75 2 0 1 1
0
243 18.78 3.00 2 1 0 1
0

day_Fri day_Sat day_Sun day_Thur time_Dinner time_Lunch

0 0 0 1 0 1 0
1 0 0 1 0 1 0
2 0 0 1 0 1 0
3 0 0 1 0 1 0
4 0 0 1 0 1 0
.. ... ... ... ... ... ...
239 0 1 0 0 1 0
240 0 1 0 0 1 0
241 0 1 0 0 1 0
242 0 1 0 0 1 0
243 0 0 0 1 1 0

[244 rows x 13 columns]

X=df_encoded.drop('tip',axis=1)

y=df_encoded['tip'].astype('int')

from sklearn.feature_selection import chi2

chi2_value,p_value =chi2(X,y)

chi2_value

array([451.45173551, 23.98377395, 1.99590599, 1.1060116 ,

2.41410127, 3.9196698 , 4.64580879, 5.68446664,
7.50191554, 8.85006698, 3.79699815, 9.82752462])

p_value

array([1.80666902e-92, 2.30619124e-03, 9.81137095e-01, 9.97485792e-01,

9.65616197e-01, 8.64297389e-01, 7.94674362e-01, 6.82528082e-01,
4.83569431e-01, 3.55101700e-01, 8.74958733e-01, 2.77340784e-
01])

import numpy as np

p_value = np.around(p_value,6)

temp =
pd.DataFrame({'features':X.columns,'chi_square':chi2_value,'p_value':p
_value})
temp

features chi_square p_value

0 total_bill 451.451736 0.000000
1 size 23.983774 0.002306
2 sex_Female 1.995906 0.981137
3 sex_Male 1.106012 0.997486
4 smoker_No 2.414101 0.965616
5 smoker_Yes 3.919670 0.864297
6 day_Fri 4.645809 0.794674
7 day_Sat 5.684467 0.682528
8 day_Sun 7.501916 0.483569
9 day_Thur 8.850067 0.355102
10 time_Dinner 3.796998 0.874959
11 time_Lunch 9.827525 0.277341

temp.sort_values('chi_square',ascending=False)

features chi_square p_value

0 total_bill 451.451736 0.000000
1 size 23.983774 0.002306
11 time_Lunch 9.827525 0.277341
9 day_Thur 8.850067 0.355102
8 day_Sun 7.501916 0.483569
7 day_Sat 5.684467 0.682528
6 day_Fri 4.645809 0.794674
5 smoker_Yes 3.919670 0.864297
10 time_Dinner 3.796998 0.874959
4 smoker_No 2.414101 0.965616
2 sex_Female 1.995906 0.981137
3 sex_Male 1.106012 0.997486

Let's walk through an example of a Chi-square test with real data. Imagine we are analyzing the
relationship between gender and preference for a new product. We want to determine whether
gender (male/female) affects the likelihood of liking or disliking the product. This creates two
categorical variables: Gender and Product Preference.

Data Example:
Gender Likes Product Dislikes Product Total
Male 30 10 40
Female 20 30 50
Total 50 40 90

Hypotheses:
• Null Hypothesis (H₀): Gender and product preference are independent (no association).
• Alternative Hypothesis (H₁): Gender and product preference are dependent (there is an
association).
Step-by-Step Chi-square Test:
1. Observed Frequencies:
The table above shows the observed frequencies (real counts from our data).

2. Expected Frequencies:
We calculate the expected frequency for each cell assuming that the two variables are
independent, using the formula:

[ E_{ij} = \frac{{\text{{Row Total}}_i \times \text{{Column Total}}_j}}{\text{{Grand Total}}} ]

For example, the expected frequency of males who like the product is:

[ E_{\text{Male, Likes}} = \frac{{40 \times 50}}{90} = 22.22 ]

Similarly, we can calculate the expected frequency for each cell:

Dislikes Product
Gender Likes Product (Expected) (Expected)
Male 22.22 17.78
Female 27.78 22.22

3. Chi-square Statistic:
We compute the Chi-square statistic using the formula:

[ \chi^2 = \sum \frac{{(O_{ij} - E_{ij})^2}}{E_{ij}} ]

Where (O_{ij}) is the observed frequency and (E_{ij}) is the expected frequency.

For each cell:

• For males who like the product: (\frac{{(30 - 22.22)^2}}{22.22} = 2.72)

• For males who dislike the product: (\frac{{(10 - 17.78)^2}}{17.78} = 3.4)
• For females who like the product: (\frac{{(20 - 27.78)^2}}{27.78} = 2.18)
• For females who dislike the product: (\frac{{(30 - 22.22)^2}}{22.22} = 2.72)

The total Chi-square statistic is:

[ \chi^2 = 2.72 + 3.4 + 2.18 + 2.72 = 11.02 ]

4. Degrees of Freedom:
The degrees of freedom for a Chi-square test is calculated as:

[ \text{{Degrees of Freedom}} = (r - 1) \times (c - 1) ]

Where:

• (r) is the number of rows (2 for male and female).

• (c) is the number of columns (2 for like and dislike).
[ \text{{Degrees of Freedom}} = (2 - 1) \times (2 - 1) = 1 ]

5. p-value:
Using the Chi-square distribution table or a calculator, we find the p-value for (\chi^2 = 11.02)
and 1 degree of freedom.

The p-value turns out to be p = 0.0009.

6. Conclusion:
Since p < 0.05, we reject the null hypothesis and conclude that gender and product preference
are dependent — there is a significant relationship between gender and liking/disliking the
product.

This is how you would apply a Chi-square test in practice!