Project 2 - EDA on Vehicle Insurance Customer Data

Project 2 – EDA on Vehicle Insurance Customer Data"

🔹 Step 1: Import & Add Column Names

python

import pandas as pd
import numpy as np

# Load your datasets
# customer_df = pd.read_csv('customer.csv')
# policy_df = pd.read_csv('customer_policy.csv')

# Add proper column names
customer_df.columns = [
    'customer_id', 'Gender', 'age', 'driving_licence_present',
    'region_code', 'previously_insured', 'vehicle_age', 'vehicle_damage'
]

policy_df.columns = [
    'customer_id', 'annual_premium', 'sales_channel_code',
    'vintage', 'response'
]

🔹 Step 2: Data Quality Checks & Cleaning

2.i Handling Nulls

python

# Null summary
print(customer_df.isnull().sum())
print(policy_df.isnull().sum())

# Drop rows with null customer_id
customer_df.dropna(subset=['customer_id'], inplace=True)
policy_df.dropna(subset=['customer_id'], inplace=True)

# Fill numeric nulls with mean
customer_df.fillna(customer_df.select_dtypes(include='number').mean(), inplace=True)
policy_df.fillna(policy_df.select_dtypes(include='number').mean(), inplace=True)

# Fill categorical nulls with mode
for df in [customer_df, policy_df]:
    for col in df.select_dtypes(include='object'):
        df[col].fillna(df[col].mode()[0], inplace=True)

2.ii Outlier Handling (IQR)

python

def treat_outliers(df):
    for col in df.select_dtypes(include='number'):
        Q1 = df[col].quantile(0.25)
        Q3 = df[col].quantile(0.75)
        IQR = Q3 - Q1
        lower = Q1 - 1.5 * IQR
        upper = Q3 + 1.5 * IQR
        df.loc[(df[col] < lower) | (df[col] > upper), col] = df[col].mean()

treat_outliers(customer_df)
treat_outliers(policy_df)

2.iii Remove Whitespaces & Case Correction

python

def clean_text_cols(df):
    for col in df.select_dtypes(include='object'):
        df[col] = df[col].str.strip().str.lower()

clean_text_cols(customer_df)
clean_text_cols(policy_df)

2.iv Dummies (optional for modeling)

python

# customer_df = pd.get_dummies(customer_df, drop_first=True)
# policy_df = pd.get_dummies(policy_df, drop_first=True)

2.v Drop Duplicates

python
customer_df.drop_duplicates(inplace=True)
policy_df.drop_duplicates(inplace=True)

🔹 Step 3: Create Master Table

python

master_df = pd.merge(customer_df, policy_df, on='customer_id', how='inner')
master_df.head()

🔹 Step 4: Business Insights

4.i Gender-wise Avg Premium

python

master_df.groupby('Gender')['annual_premium'].mean()

4.ii Age-wise Avg Premium

python

master_df.groupby('age')['annual_premium'].mean()

4.iii Gender Balance Check

python

master_df['Gender'].value_counts()

4.iv Vehicle Age-wise Avg Premium

python

master_df.groupby('vehicle_age')['annual_premium'].mean()

🔹 Step 5: Correlation Between Age & Annual Premium

python

correlation = master_df['age'].corr(master_df['annual_premium'])
print("Correlation:", correlation)

# Interpretation
if correlation > 0.5:
    print("Strong positive relationship")
elif correlation < -0.5:
    print("Strong negative relationship")
else:
    print("No significant relationship")