Untitled

import pandas as pd
from matplotlib import pyplot as plt
import seaborn as sns

#Statsmodels
import statsmodels.api as sm

#Scikit learn
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
#Before we imported LinearRegression from scikit learn, now we import LogisticRegression


df1.isnull().any()

df2.describe()

df1["Gender_dummy"] = df1["Gender"].replace({"Male":0, "Female":1})

df2 = df1.drop("Gender", axis = 1)

correlations = df2.corr()
sns.heatmap(correlations, annot = True).set(title = "Heatmap of Consumption Data - Pearson Correlations")

y = df2["Purchased"]
X = df2[["Gender_dummy", "Age", "AnnualSalary"]]

X = sm.add_constant(X)
log_reg = sm.Logit(y,X, data=df2).fit()
log_reg.summary()

import numpy as np

odds_ratios = pd.DataFrame(
    {
        "OR": log_reg.params,
    }
)
odds_ratios = np.exp(odds_ratios)

print(odds_ratios)