EDA on Vehicle Insurance Customer Data

EDA on Vehicle Insurance Customer Data

Introduction:

The purpose of this project is to analyze customer data to gather insights and make data-driven decisions for the company. The customer data is divided into two tables: customer details and customer_policy. The Data Science team is tasked with performing various data quality checks, data cleaning, merging the tables, and deriving important information from the master table.

Tasks:

The Data Science team has identified the following tasks to be performed:

1. Add Column Names: i. Customer Details Table:

• customer_id
• Gender
• age
• driving_license_present
• region_code
• previously_insured
• vehicle_age
• vehicle_damage

ii. Customer Policy Table:

• customer_id
• annual_premium (in Rs)
• sales_channel_code
• vintage
• response

2. Checking and Cleaning Data Quality: i. Null Values:

• Generate a summary of the count of all null values column-wise.
• Drop null values for customer_id because central tendencies for IDs are not feasible.
• Replace null values for numeric columns with the mean.
• Replace null values for categorical columns with the mode.

ii. Outliers:

• Generate a summary of the count of all outliers column-wise.
• Replace outlier values for numeric columns with the mean.
• Identify outliers using the IQR method and replace values below Q1 - 1.5IQR or above Q3 + 1.5IQR.

iii. White Spaces:

• Remove white spaces from the data.

iv. Case Correction:

• Convert the data to either lower case or upper case.

v. Convert Nominal Data into Dummies:

• Convert categorical variables into dummy variables for future modelling if required.

vi. Drop Duplicates:

• Remove duplicated rows from the data.

3. Create a Master Table: Join the customer details table and customer_policy table using customer_id to create a master table for future use.

Important Information from the Master Table:

The company requires the following information from the master table for decision-making:

i. Gender-wise average annual premium.

ii. Age-wise average annual premium.

iii. Checking if the data is balanced between genders (approximate count equality).

iv. Vehicle age-wise average annual premium.

Relationship Between Person Age and Annual Premium:

Analyze the correlation between a person's age and annual premium using the correlation coefficient. Determine the strength of the relationship based on the coefficient value.

Conclusion:

The project involves analyzing customer data to gain insights and provide valuable information for the company's decision-making process. By performing data quality checks, cleaning, merging, and analysis, the Data Science team aims to provide actionable recommendations based on the findings.