Set Up

Vehicle insurance also known as auto insurance, car insurance or motor insurance is a contract between the insured (vehicle owner) and an insurance company (insurer) which provides protection financially in the event of an accident or theft that may result in vehicle damages or bodily injuries.

Exploratory Data Analysis (EDA) will be performed in order to understand certain pattern within the insurance company activities and customers/policyholders preference. Some important features from the data that make up part of the various activities in auto insurance such as coverage type, premium payment, type of policy, customer claim amount, etc will be looked into.

R will be used for the analysis and the various custom functions used in the analysis can be seen in [insurance_function.R]

Functions and Libraries

source("eda_functions.R", local = knitr::knit_global())

library(tidyverse) |> suppressPackageStartupMessages()
library(patchwork) 
library(tidytext, include.only = c("reorder_within", "scale_y_reordered"))
library(janitor, include.only = "clean_names")

Data

data <- read_csv("Marketing-Customer-Value.csv")
 
# Data cleaning.
data <- data %>% 
  mutate(Gender = case_when(Gender == "F" ~ "Female",
                            Gender == "M" ~ "Male", TRUE ~ Gender),
        `Vehicle Size` = if_else(`Vehicle Size` == "Medsize", 
                                 "Midsize", `Vehicle Size`)) %>% 
  select(- c(Customer, `Renew Offer Type`, `Effective To Date`)) %>% 
  clean_names()

Data Dictionary

• CLV: This is the monetary value of a customer relationship based on the present value of the projected future cash flows.
  
• monthly premium auto: This is the amount paid to the insurance company on a monthly basis, in exchange for insurance coverage.
  
• months since last claim: The number of months since the customer last made a claim.
  
• months since policy inception: The number of months since the policy was initiated.
  
• number of policies: The number of policy owned by a policyholder.
  
• number of open complaints: Open complains made by a policyholder.
  
• total claim amount: Total amount paid to policyholders for the loss covered in their policy.
  
• coverage: The type of damages, theft and others covered by the insurance policy.
  
• education: The customer level of education.
  
• state: The state which the customer resides.
  
• response: Customer response to the last campaign.
  
• employment status: Customer employment status.
  
• gender: The sex of a customer.
  
• location code: The type of area or community in which a customer resides.
  
• marital status: Customer’s Marital or relationship status.
  
• policy type: The types of policy owned by the policyholder.
  
• sales channel: The channel of through which customers were acquired.
• vehicle class: The type of the vehicle class insured.
  
• vehicle size : The size of the vehicle insured.
  
• income: Customer income.

Questions to answer from the data.

1. Analysis on Customer Lifetime Values
   What is the average CLV based on each channel of customer acquisition.
   Did customer income have an affect on its CLV.
   What is the relationship between CLV and number of policy owned.
   What are the locations of the company’s most valuable customers.
2. Various Features Used In Determining Auto Premium
   What are the relationship between the features used in determining vehicle premium in the data.
   
3. Analysis On Number Of Policies Owned By policyholders
   How did customers with multiple policies fare with total amount claimed and how much open complains were made.
   What is the relationship between number of policy owned by a policyholder and the type of policy.
   
4. Sales Channels And Vehicle Classification
   Which vehicle class and size have the Lowest total amount claimed.
   What are the types of policies that were initiated through each sales channel.
   Sale channel performance by customer location. How did each sales channel performed in each location.
   
5. Analysis On Types of Auto Policy and Customer Income
   What is the Number of customers that responded to the last campaign by each type of policy.
   How did policy type change for each gender and type of coverage.
   Did income have a contributing effect on the type of coverage, class & size of vehicles.