Hospital-Readmission-Analysis

🏥 Hospital Readmission Analysis

An end-to-end analytics and modeling solution designed to assess patient readmission risk across specialties and treatment patterns. This project analyzes 25,000+ patient records to empower hospitals with predictive insights, operational dashboards, and data-driven care strategies.

GitHub Repository

🔗 click to view Hospital-Readmission-Analysis

🧠 Project Overview

Hospital readmissions are a key metric for care quality and operational efficiency. This project delivers a full-stack solution that enables:

🔍 Risk profiling of patients based on visit history and diagnosis
📊 Specialty-level readmission trend analysis
🧠 Predictive modeling for care optimization
📈 Dashboard-driven insights for hospital decision-making

🎯 Key Objectives

Clean and preprocess patient-level hospital data
Engineer features for readmission prediction and dashboarding
Build classification models to assess readmission risk
Deploy interactive dashboards for real-time clinical insights

📁 Project Structure

File Name	Description
`hospital_readmission.sql`	SQL queries for data extraction and transformation
`hospital readmission.ipynb`	Jupyter notebook with full analysis workflow
`sqlconnect.py`	Python script for SQL database connection
`app.py`	Streamlit app for interactive model deployment
`readmission_model.pkl`	Trained classification model for readmission prediction
`feature_names.pkl`	Serialized feature list used in model training
`cleaned_hospital_readmission.csv`	Preprocessed dataset used for modeling
`hospital_readmissions_cleaned.csv`	Alternate cleaned dataset version
`Hospital Readmission Analytics.docx`	Project documentation and summary report

📊 Dataset Summary

Total Records: 25,000
Key Columns:
age, time_in_hospital, n_lab_procedures, n_procedures, n_medications, n_outpatient, n_inpatient, n_emergency, medical_specialty, diag_1, diag_2, diag_3, glucose_test, A1Ctest, change, diabetes_med, readmitted, total_visits
Target Variable: readmitted
Feature Set: Visit counts, diagnosis codes, medication history, specialty, and encoded categorical variables

🧹 Data Preprocessing

Converted date fields and calculated total visits
Normalized treatment metrics and encoded categorical features
Removed outliers and handled missing values
Engineered features for diagnosis grouping and specialty impact

📈 Exploratory Data Analysis

📊 Readmission trends by specialty and diagnosis
🏥 Visit patterns across inpatient, outpatient, and emergency channels
💊 Medication change and diabetes treatment impact
📅 Hospital stay duration and seasonal readmission patterns

🤖 Modeling Approach

Target Variable: readmitted
Algorithms Used: Logistic Regression, Random Forest, XGBoost
Evaluation Metrics: Accuracy, Precision, Recall, F1 Score, ROC-AUC
Top Features: n_inpatient, change, medical_specialty, diag_1, n_medications