Healthcare Analytics Project

End-to-end analysis of 10,000 healthcare encounters with predictive modeling and patient segmentation.

Quick Summary

• Dataset: 10,000 encounters, 15 dimensions, Synthetic. Source -> here
• Models: Logistic Regression & Random Forest
• Segmentation: 6 patient clusters identified with distinct profiles
• Revenue: $255.2M analyzed with clear cost drivers

Get Started

pip install -r requirements.txt

# View the visualization dashboard
jupyter notebook Concise-summary.ipynb

# Or run notebooks in order
jupyter notebook notebooks/

Detailed Reports

• 📊 DATA_ANALYSIS_REPORT.md – Complete EDA with statistics
• 🤖 MODEL_RESULTS_REPORT.md – Model performance & cluster insights
• 📈 Concise-summary.ipynb – Interactive visualization viewer

Key Findings

Metric	Result
Risk Model Accuracy	100% (test set)
Clusters Identified	6 distinct patient cohorts
Data Quality	99.9% valid
Mean Billing	$25,516.81
Mean LOS	15.6 days

Project Structure

notebooks/                          – Analysis pipeline
├── 00_environment_setup.ipynb      – Data loading
├── 01_eda_raw_tables.ipynb         – Exploratory analysis
├── 02_sql_validation_and_joins.sql – Data warehouse
└── 03_advanced_analysis_and_modeling.ipynb – Models

data/
├── raw/                            – Original CSV
└── processed/                       – Cleaned exports

docs/                               – Data dictionary, SQL reference
healthcare.db                       – SQLite database
Concise-summary.ipynb                     – Visualization viewer
DATA_ANALYSIS_REPORT.md             – EDA report
MODEL_RESULTS_REPORT.md             – Modeling report

Cluster Profiles (6 Segments)

Cluster	Size	Age	LOS	Billing	Top Condition
0: High-Acuity Elderly	16%	61.2	23.3 days	$39,467	Diabetes
1: Young/Acute	16%	40.8	8.2 days	$12,340	Asthma
2: Moderate Mixed	17%	52.1	15.6 days	$25,218	Hypertension
3: Younger Chronic	17%	48.3	14.1 days	$22,650	Arthritis
4: High-Complexity	17%	55.8	21.3 days	$34,120	Cancer
5: Quick-Resolution	16%	45.2	10.1 days	$14,580	Obesity

More Details

1. MODEL_RESULTS_REPORT.md
2. Concise-summary.ipynb