Project 2 - EDA on Vehicle Insurance Customer Data

In this project, the Data Science team was tasked with performing various data preparation and analysis tasks on two datasets: one containing customer details and the other containing policy details for the company's customers. The project involved several key steps and data quality checks to extract meaningful insights for future decision-making. Below is a summary of the project's key findings and conclusions:

1. Data Preparation:

   • Column Names: Column names were added to both datasets, ensuring clarity and consistency.
   • Customer Details Table:
     • Columns: The columns were renamed to 'customer_id,' 'Gender,' 'age,' 'driving licence present,' 'region code,' 'previously insured,' 'vehicle age,' and 'vehicle damage.'
   • Customer Policy Table:
     • Columns: The columns were renamed to 'customer_id,' 'annual premium (in Rs),' 'sales channel code,' 'vintage,' and 'response.'

2. Data Quality Checks:

   • Null Values:

     • A summary of null values was generated for both datasets to identify missing data.
     • Null values in the 'customer_id' column were dropped as customer IDs are identifiers.
     • Null values in numeric columns were replaced by their mean values.
     • Null values in categorical columns were replaced by their mode values.

   • Outliers:

     • A summary of outlier counts was generated using the IQR method.
     • Outliers in numeric columns were replaced by mean values.

   • White Spaces:

     • White spaces in the data were removed.

   • Case Correction:

     • Data was converted to a consistent case (lower or upper case).

   • Convert Nominal Data to Dummies:

     • Categorical data was converted into dummy variables for future modeling.

   • Duplicates:

     • Duplicate rows were removed from the datasets.

3. Master Table Creation:

   • A master table was created by joining the 'customer' and 'customer_policy' tables using 'customer_id' as the common key. This master table will be useful for future analysis and modeling.

4. Key Information for Decision-Making:

   • Gender-Wise Average Annual Premium:

     • The average annual premium was calculated separately for each gender.

   • Age-Wise Average Annual Premium:

     • The average annual premium was calculated for different age groups.

   • Gender Balance:

     • The data was checked to determine whether it's balanced between genders by comparing the counts in each group.

   • Vehicle Age-Wise Average Annual Premium:

     • The average annual premium was calculated for different vehicle age categories.

5. Correlation Analysis:

   • A correlation analysis was performed to determine the relationship between a person's age and the annual premium.
   • Correlation coefficients were used to describe the strength and direction of the relationship.

In conclusion, this project successfully prepared and analyzed the customer and policy data to provide valuable insights for the company. The cleaned and structured data can now be used for future decision-making, including pricing strategies, targeting specific customer segments, and assessing the correlation between age and premium. This data-driven approach will aid the company in making informed business decisions and optimizing its operations for future growth