Insurance Analytics Dashboard

Overview

Rows

278,860

Columns

Missing Values

0 total

Duplicate Rows

210.71 MB in memory

Data Quality

Missing Values by Column

No missing values detected

Outlier Summary

No outliers detected (IQR method)

Distributions

Age — Distribution

Age — Box Plot

Min: 18.00Q1: 30.00Med: 41.00Q3: 53.00Max: 64.00

Annual Income — Distribution

Annual Income — Box Plot

Min: 0.00Q1: 14412.00Med: 32191.00Q3: 60028.75Max: 128453.88

Number of Dependents — Distribution

Number of Dependents — Box Plot

Min: 0.00Q1: 1.00Med: 2.00Q3: 3.00Max: 4.00

Health Score — Distribution

Health Score — Box Plot

Min: 0.04Q1: 16.55Med: 26.45Q3: 38.36Max: 71.08

Previous Claims — Distribution

Previous Claims — Box Plot

Min: 0.00Q1: 0.00Med: 1.00Q3: 1.00Max: 2.50

Vehicle Age — Distribution

Vehicle Age — Box Plot

Min: 0.00Q1: 5.00Med: 10.00Q3: 15.00Max: 19.00

Credit Score — Distribution

Credit Score — Box Plot

Min: 300.00Q1: 452.00Med: 575.00Q3: 697.00Max: 849.00

Insurance Duration — Distribution

Insurance Duration — Box Plot

Min: 1.00Q1: 3.00Med: 5.00Q3: 7.00Max: 9.00

Premium Amount — Distribution

Premium Amount — Box Plot

Min: 0.00Q1: 288.00Med: 688.00Q3: 1360.00Max: 2968.00

Summary Statistics

Column	Count	Mean	Std	Min	Median	Max	Skew	Kurt
Age	278,860	41.02	13.44	18.00	41.00	64.00	-0.00	-1.17
Annual Income	278,860	41321.80	33867.38	0.00	32191.00	128453.88	0.97	0.10
Number of Dependents	278,860	2.00	1.34	0.00	2.00	4.00	0.00	-1.11
Health Score	278,860	28.46	15.53	0.04	26.45	71.08	0.57	-0.21
Previous Claims	278,860	0.95	0.72	0.00	1.00	2.50	0.36	-0.38
Vehicle Age	278,860	9.52	5.77	0.00	10.00	19.00	-0.00	-1.21
Credit Score	278,860	574.43	150.64	300.00	575.00	849.00	0.00	-1.00
Insurance Duration	278,860	5.01	2.58	1.00	5.00	9.00	-0.00	-1.23
Premium Amount	278,860	933.49	814.70	0.00	688.00	2968.00	1.05	0.20

Correlations

Pearson Correlation Matrix

-1

Top Correlated Pairs

Variable 1	Variable 2	Correlation
Credit Score	Premium Amount	0.0045
Insurance Duration	Premium Amount	0.0037
Annual Income	Credit Score	-0.0032
Annual Income	Number of Dependents	0.0030
Health Score	Previous Claims	0.0030
Age	Premium Amount	0.0023
Annual Income	Previous Claims	-0.0023
Number of Dependents	Health Score	0.0023
Health Score	Vehicle Age	-0.0022
Number of Dependents	Credit Score	-0.0020

Cramer's V (Categorical Associations)

Variable 1	Variable 2	Cramer's V
Smoking Status	Risk Category	0.9429
Exercise Frequency	Risk Category	0.2884
Smoking Status	Exercise Frequency	0.0053
Age Group	Risk Category	0.0049
Marital Status	Age Group	0.0047
Education Level	Occupation	0.0047
Gender	Property Type	0.0045
Marital Status	Income Bracket	0.0045
Education Level	Age Group	0.0043
Income Bracket	Risk Category	0.0043

Data Preprocessing

Missing Values Filled

272,794

Outliers Capped

38,186

Features Engineered

Imputation Strategies

Missing Values: Before vs After

Imputation Log

Column	Strategy	Fill Value	Count
Age	median	41	4,685
Annual Income	median	32,191	13,955
Number of Dependents	median	2	27,886
Health Score	median	26.451	10,597
Previous Claims	median	1	81,288
Credit Score	median	575	27,886
Premium Amount	median	688	1,841
Marital Status	mode	Single	5,019
Occupation	constant (Unknown)	Unknown	81,288
Customer Feedback	constant (Unknown)	Unknown	18,349

Engineered Features

Age Group5 unique values

Binned into 18-25/26-35/36-45/46-55/56-65/65+

Income Bracket4 unique values

Quartile-based: Low/Medium/High/Very High

Risk Category3 unique values

Composite of Smoking + Exercise + Health Score

Outlier Treatment (IQR Capping)

Annual Income7,553 capped [-54,013.125 ~ 128,453.875]

Health Score2,457 capped [-16.16 ~ 71.076]

Previous Claims15,871 capped [-1.5 ~ 2.5]

Premium Amount12,305 capped [-1,320 ~ 2,968]

Premium Prediction Model

Model

Ridge Regression

R-squared

-0.0003

Higher is better

RMSE

814.64

Lower is better

MAE

658.59

Lower is better

Test Size

55,772

Model Comparison (5-Fold Cross-Validation)

Model	R²	RMSE	MAE	CV Mean (R²)	CV Std
Random Forest	-0.0042	816.22	660.14	-0.0080	0.0016
Gradient Boosting	-0.0132	819.86	662.17	-0.0223	0.0044
XGBoost	-0.0633	839.9	675.46	-0.0838	0.0076
Ridge RegressionBest	-0.0003	814.64	658.59	-0.0016	0.0010

Feature Importance (Top 15)

Actual vs Predicted

Residual Distribution

Customer Segmentation

Method

K-Means

Clusters

Features Used

PCA Var Explained

33.9%

Elbow Method

Cluster Distribution

PCA 2D Visualization

Cluster Profiles

Cluster	Size	Age	Annual Income	Health Score	Credit Score	Premium Amount	Insurance Duration
Young Budget	9,665 (32.2%)	29.03	27,001.83	27.69	567.67	591.54	4.93
Mid-Age Standard	5,425 (18.1%)	40.44	98,065.92	28.4	576.98	767.53	5.2
Senior Premium	5,310 (17.7%)	41	33,647.98	28.58	579.38	2,335.49	5.04
High-Value	9,600 (32%)	53.4	27,909.26	29.22	578.42	600.51	4.97

Hypothesis Testing

Smokers vs Non-Smokers Premium

Not Significant

Smokers

935.4

n=139,635

Non-Smokers

931.58

n=139,225

Test: Welch's T-Test | t = 1.2405 | p = 0.214795

Smokers have 0.4% higher premiums than Non-Smokers (p=0.2148, not significant)

High vs Low Health Score Premium

Not Significant

Health Score >= 26.5

935.23

n=144,729

Health Score < 26.5

931.62

n=134,131

Test: Welch's T-Test | t = 1.1665 | p = 0.243426

No significant difference in premiums between high and low health score groups (p=0.2434)

Comprehensive vs Premium Policy Premium

Not Significant

Comprehensive

934.35

n=92,479

Premium

931.58

n=93,298

Test: Welch's T-Test | t = 0.7351 | p = 0.462256

Comprehensive have 0.3% higher premiums than Premium (p=0.4623, not significant)

Customer Default Rate Prediction

Overall Default Rate

73.9%

Best Model

Random Forest

Train Size

16,000

Test Size

4,000

Model Comparison (5-Fold CV)

Model	Accuracy	Precision	Recall	F1	CV Mean	CV Std
Random ForestBest	74.2%	74.2%	100.0%	85.2%	74.2%	0.0001
Gradient Boosting	73.9%	74.2%	99.4%	84.9%	73.9%	0.0020
XGBoost	73.0%	74.5%	96.9%	84.2%	72.5%	0.0025
Logistic Regression	74.2%	74.2%	100.0%	85.2%	74.2%	0.0001

Overview

Data Quality

Missing Values by Column

Outlier Summary

Distributions

Age — Distribution

Age — Box Plot

Annual Income — Distribution

Annual Income — Box Plot

Number of Dependents — Distribution

Number of Dependents — Box Plot

Health Score — Distribution

Health Score — Box Plot

Previous Claims — Distribution

Previous Claims — Box Plot

Vehicle Age — Distribution

Vehicle Age — Box Plot

Credit Score — Distribution

Credit Score — Box Plot

Insurance Duration — Distribution

Insurance Duration — Box Plot

Premium Amount — Distribution

Premium Amount — Box Plot

Summary Statistics

Correlations

Pearson Correlation Matrix

Top Correlated Pairs

Cramer's V (Categorical Associations)

Data Preprocessing

Missing Values: Before vs After

Imputation Log

Engineered Features

Outlier Treatment (IQR Capping)

Premium Prediction Model

Model Comparison (5-Fold Cross-Validation)

Feature Importance (Top 15)

Actual vs Predicted

Residual Distribution

Customer Segmentation

Elbow Method

Cluster Distribution

PCA 2D Visualization

Cluster Profiles

Hypothesis Testing

Smokers vs Non-Smokers Premium

High vs Low Health Score Premium

Comprehensive vs Premium Policy Premium

Customer Default Rate Prediction

Model Comparison (5-Fold CV)

Default Rate by Age Group

Default Risk Factors