EDA

eda_1.py

from sklearn.datasets import load_diabetes
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
# LOADING DIABETES DATA (INPUT FEATURES) AND STORING IT IN A DATA FRAME
data = pd.DataFrame(load_diabetes()["data"],columns=load_diabetes()["feature_names"])
#ADDING TARGET VARIABLE TO THE DATA FRAME
data["target"] = load_diabetes()["target"]

eda_10.py

numeric_features = list(data.select_dtypes("float64").columns)
numeric_features.remove('target')
categorical_features = list(data.select_dtypes("int8").columns)
target = "target"
print(f'numeric_features:\n{numeric_features}\n\ncategorical_features:\n{categorical_features}\n\ntarget:\n{target}')

eda_11.py

np.round(data.isnull().mean() * 100,1)

eda_12.py

for i in categorical_features:
    print(f'{i}\n{np.round((data[i].value_counts() / data[i].value_counts().sum()) * 100,2)}')

eda_13.py

fig,ax = plt.subplots(3,3,figsize=(10,10))
row = col = 0
for n,i in enumerate(numeric_features):
    if (n%3 == 0) & (n > 0):
        row += 1
        col = 0
    data[i].plot(kind="kde",ax=ax[row,col])
    ax[row,col].set_title(i)
    col += 1

eda_14.py

from scipy.stats import normaltest
for i in numeric_features:
    print(f'{i}: {"Not Gaussian" if normaltest(data[i].values,)[1]<0.05 else "Gaussian"}  {normaltest(data[i].values)}')

eda_15.py

for i in numeric_features:
    print(f'{i}: {np.abs(np.round((data[i].std()/data[i].median()) * 100,2))}')

eda_16.py

fig,ax = plt.subplots(1,3,figsize=(20,5))
data[target].plot(kind="hist",ax=ax[0])
data[target].plot(kind="kde",ax=ax[1])
data[target].plot(kind="box",ax=ax[2])
plt.show()
print(f'{target}: {"Not Gaussian" if normaltest(data[target].values,)[1]<0.05 else "Gaussian"}  {normaltest(data[target].values)}')

eda_17.py

sns.boxplot(x=data[categorical_features[0]],y=data[target])

eda_18.py

fig,ax = plt.subplots(3,3,figsize=(15,10))
row = col = 0
for n,i in enumerate(numeric_features):
    if (n%3 == 0) & (n > 0):
        row += 1
        col = 0
    sns.regplot(x=i,y="target",data=data,ax=ax[row,col],ci=False)
    col += 1

eda_19.py

num_tgt = numeric_features.copy() 
num_tgt.append('target')
fig = plt.figure(figsize=(8,8))
sns.heatmap(data[num_tgt].corr(method='pearson'),annot=True,fmt='.2f',mask=np.triu(data[num_tgt].corr(method='pearson')),cbar=False)

eda_2.py

data.shape

eda_20.py

fig = plt.figure(figsize=(8,8))
sns.heatmap(data[num_tgt].corr(method='kendall'),annot=True,fmt='.2f',mask=np.triu(data[num_tgt].corr(method='pearson')),cbar=False)

eda_3.py

data.head()

eda_4.py

list(data.columns)

eda_5.py

data.dtypes

eda_6.py

data["sex"].unique()

eda_7.py

data.loc[data["sex"]>0,"sex"] = 1
data.loc[data["sex"]<0,"sex"] = 0
data.head()

eda_8.py

data.describe().T

eda_9.py

data["sex"] = data["sex"].astype(np.int8)
data.dtypes