Introduction

For this Module the data set I chose is Auto Insurance Claims, which is a dataset containing information about claims made by customers and the corresponding insurance information.

Dataset

The data has 26 columns and 9135 rows. The Columns consist of Customer, Country, State Code, State, Claim Amount, Response, Coverage, Education, Effective To Date, Employment Status, Gender, Income, Location Code, Marital Status, Monthly Premium Auto, Months Since Last Claim, Months Since Policy Inception, Number of Open Complaints, Number of Policies, Policy Type, Policy, Claim Reason, Sales Channel, Total Claim Amount, Vehicle Class, and Vehicle Size. Each row is a customer and each customer had data for the 25 other data fields. The data is all from the United States, which is denoted by “US” in “Country”. It also is data from only five states, Kansas (KS), Nebraska (NE), Oklahoma (OK), Missouri (MO), and Iowa (IA). Each customer is categorized by gender, marital status, income, education, and vehicle size. The insurance details include everything else. One limitation of the data is the “Effective to Date” variable. It describes the day that the policy given ends, rather than claimed. While interesting, a more valuable date could have been date filed and that could be used to gather interesting data of patterns of what days more claims are filed. Overall, this dataset provides an opportunity to examine the relationships between customer characteristics, insurance claims, vehicles, and claim types.

Findings

Not many of the rows have a numerical value but of those who do we can calculate the Sum, Median, Mean, Minimum, and Maximum.

I ran through each of those functions for all the columns consisting of numerical values. This includes, Claim Amount, Income, Total Claim Amount, Months Premium Auto, Months Since Last Claim, Months Since Policy Inception, Number of Open Complains, and Number of Policies.

For Claim Amount: Sum = 7311712.63, Median = 578.01822, Mean = 800.49405, Minimum = 189.80076, Maximum = 8332.538

For Income: Sum = 343962509, Median = 33889.5, Mean = 37657.38001, Minimum = 0, Maximum = 99981

For Total Claim Amount: Sum = 3964967.047, Median = 383.9454, Mean = 434.0888, Minimum = 0.099007, Maximum = 2893.23968

For Months Premium Auto: Sum = 851465, Median = 83, Mean = 93.2192906, Minimum = 61, Maximum = 298

For Months Since Last Claim: Sum = 137896, Median = 14, Mean = 15.0970002, Minimum = 0, Maximum = 35

For Months Since Policy Inception: Sum = 439022, Median = 48, Mean = 48.0646, Minimum = 0, Maximum = 99

For Number of Open Complains: Sum = 3511, Median = 0, Mean = 0.384388, Minimum = 0, Maximum = 5

And final for Number of Policies: Sum = 27093, Median = 2, Mean = 2.96612, Minimum = 1, Maximum = 9

While this data is helpful to understand ideas about Income and Claim Amounts, by looking at more descriptive visualizations we can figure out more about relationships between each variable.

Tab 1

The first graph is a layered bar graph, Claim Reason By Each State. This graph looks at each state in the data set and how many claims there are. Then within each state the number of different types of data is denoted by a different color. This allows us to easily compare the overall number of claims and the composition of those claims numerically. This allows us to see if there are any patterns for each state, and what differences there are.

Tab 2

The second graph is pie charts, Percentage of Claims Reasons Within Each State. These charts follow a similar idea to graph 1. It looks at the percentage of each claim per State. This can be used to more easily compare the contribution of each claim to the overall claims for each state in the dataset.

Tab 3

The third graph is a line graph looking at the Average Claim Amount compared to the Months Since Last Claim. This allows us to see if there is any correlation between the amount of time since a customer’s last claim, and how much the customer is asking for.

Tab 4

The fourth graph is a layered line graph, Customer Count by Number of Policies and State. This looks at a similar thing to four, with how many customers each state has a certain number of policies. This allows us to see if there is a common connection between states, and how many policies are the most common to have.

Tab 5

The fifth and last graph is a Heatmap, Type of Citations by Vehicle Class. This allows us to look at if there is any relationship between the type of car and the type of Claim. This allows companies to understand if a certain Claim is more likely to be attached to a specific car, or Vehicle Class.

Conclusion

The Claim Reason by State visualization allows differences in claim types across states to be examined, while the pie charts show the relative percentage of claims represented by each state. The line graph examining Months Since Last Claim allows the relationship between previous claims and current claim amounts to be explored. The layered line graph shows how the number of policies is distributed across states, and the heatmap allows claim reasons to be compared across vehicle classes. Together, these visualizations demonstrate how descriptive analysis can be used to better understand an insurance dataset. While these graphs identify patterns and relationships, they do not establish causation. Additional information, particularly the actual date a claim was filed, could allow for a more detailed analysis of trends and potential factors influencing insurance claims.