U.S. Medical Insurance Costs

This project explores the US medical insurance dataset (available at https://www.kaggle.com/datasets/teertha/ushealthinsurancedataset). The goal is to clean, analyze, and uncover insights about factors that influence insurance charges. The analysis is structured in five steps: data loading, exploration, cleaning, analysis, and conclusions.

1. Data Loading

First, we import the necessary libraries and load the dataset into a Pandas DataFrame.

import pandas as pd

#importing and looking at the file format
insurance_data = pd.read_csv("insurance.csv")
print(insurance_data.head())

   age     sex     bmi  children smoker     region      charges
0   19  female  27.900         0    yes  southwest  16884.92400
1   18    male  33.770         1     no  southeast   1725.55230
2   28    male  33.000         3     no  southeast   4449.46200
3   33    male  22.705         0     no  northwest  21984.47061
4   32    male  28.880         0     no  northwest   3866.85520

2. Data Exploration

We start by exploring the first few rows of the dataset and checking basic info to understand the structure and types of data we are working with.

import pandas as pd

#importing and looking at the file format
insurance_data = pd.read_csv("insurance.csv")
print(insurance_data.head())

   age     sex     bmi  children smoker     region      charges
0   19  female  27.900         0    yes  southwest  16884.92400
1   18    male  33.770         1     no  southeast   1725.55230
2   28    male  33.000         3     no  southeast   4449.46200
3   33    male  22.705         0     no  northwest  21984.47061
4   32    male  28.880         0     no  northwest   3866.85520

3. Data Cleaning

We check for missing values and clean the dataset if needed. This ensures the data is ready for analysis.

# Verifying there are no empty/null values
insurance_data.isnull().sum()

age         0
sex         0
bmi         0
children    0
smoker      0
region      0
charges     0
dtype: int64

4. Data Analysis

We perform an analysis of the dataset, exploring relationships between variables such as age, BMI, smoking status, and insurance charges.

print(insurance_data.describe().round(2))

           age      bmi  children   charges
count  1338.00  1338.00   1338.00   1338.00
mean     39.21    30.66      1.09  13270.42
std      14.05     6.10      1.21  12110.01
min      18.00    15.96      0.00   1121.87
25%      27.00    26.30      0.00   4740.29
50%      39.00    30.40      1.00   9382.03
75%      51.00    34.69      2.00  16639.91
max      64.00    53.13      5.00  63770.43

# Average charges by categorical variables
print(insurance_data.groupby("sex")["charges"].mean().round(2))
print(insurance_data.groupby("smoker")["charges"].mean().round(2))
print(insurance_data.groupby("region")["charges"].mean().round(2))

sex
female    12569.58
male      13956.75
Name: charges, dtype: float64
smoker
no      8434.27
yes    32050.23
Name: charges, dtype: float64
region
northeast    13406.38
northwest    12417.58
southeast    14735.41
southwest    12346.94
Name: charges, dtype: float64

# BMI average by categorical variables
print(insurance_data.groupby("sex")["bmi"].mean().round(2))
print(insurance_data.groupby("smoker")["bmi"].mean().round(2))
print(insurance_data.groupby("region")["bmi"].mean().round(2))

sex
female    30.38
male      30.94
Name: bmi, dtype: float64
smoker
no     30.65
yes    30.71
Name: bmi, dtype: float64
region
northeast    29.17
northwest    29.20
southeast    33.36
southwest    30.60
Name: bmi, dtype: float64

# Visualization - importing libraries and creating heatmap of num correlations
import seaborn as sns
import matplotlib.pyplot as plt

correlations = insurance_data.corr(numeric_only=True)
plt.figure(figsize=(8,6))
sns.heatmap(correlations, annot=True, cmap="coolwarm", fmt=".2f")
plt.title("Correlation heatmap")
plt.show()

# Scatterplot with regression line of statistically relevant numeric variables (age and BMI)
sns.lmplot(x="age", y="charges", data=insurance_data)
plt.title("Scatterplot of Age vs Charges")
plt.savefig("images/bmi_vs_charges.png", dpi=300, bbox_inches="tight") 
plt.show()

sns.lmplot(x="bmi", y="charges", data=insurance_data)
plt.title("Scatterplot of bmi vs Charges")
plt.show()

5. Conclusions

From the analysis, we can summarize:

• Age and BMI show a correlation with insurance charges.
• Smokers tend to have significantly higher insurance costs.
• Region has less of an effect compared to lifestyle factors.

This project demonstrates how data analysis can uncover key drivers of insurance costs.