🏥

Findour HealthAI

Turning Healthcare Data into Explainable, Actionable Insights
⚕

Load Patient Encounter Dataset

Drag & drop your diabetic_data.csv file here, or click to browse.
Runs entirely in your browser — no data leaves this device, and no internet connection is required.

Ready — load a healthcare CSV dataset, or start Demo Mode
📋 50-column encounter schema
🎯 Target: readmitted (<30 days)
🧠 In-browser logistic regression
🔒 Offline · no upload · no backend
⚠️ Prototype software — not a medical device. Findour HealthAI produces statistical risk estimates to support human review and care prioritisation. It does not diagnose, treat, or make clinical decisions. All outputs require qualified human review.
FINDOUR SERVICES PRIVATE LIMITED  ·  FINDOUR SERVICES BY BHOPAL SINGH
⚠️ DEMO MODE — SYNTHETIC DATA — NOT FOR CLINICAL USE. Generated records for software evaluation only. These are not real patients.
Readmission Distribution
Admission Types
Age Distribution
Length of Stay Distribution
📈 Excel-Style Analysis
Statistical summaries, distributions & readmission breakdowns
Statistical Summary – Numerical Features
Race Distribution
Readmission by Gender (Stacked)
Readmission by Age Group (Stacked)
Readmitted Category Counts
🗄 SQL Analysis
Simulated SQL queries on the healthcare schema
🐍 Python EDA – Jupyter Notebook
Exploratory Data Analysis simulation
📉 Tableau-Style Dashboard
Interactive visualizations across all readmission dimensions
Diagnoses & Emergencies by Readmitted
Age Bucket Count by Readmitted (Stacked)
Clinical Metrics by Readmitted
Bubble Chart – Medications vs Lab Procedures (bubble size = Diagnoses, color = Readmitted)
🧠 Risk Intelligence
Prototype readmission model · risk stratification · held-out performance · patient-level explainability
🧠 Risk scores are produced by a logistic regression model trained in your browser on the currently loaded dataset, evaluated with a held-out train/test split. The prototype is a 30-Day Readmission Risk Assessment at Discharge — it uses information available by the discharge decision point, not at admission. The target readmitted is never used as an input feature. Scores are decision support for human review — not a diagnosis, and not a treatment decision.

Training model and scoring patients — please wait…

🩺 Patient Review
Human-in-the-loop review queue — every flagged patient is reviewed by a person, not by the model
👤 Human-in-the-loop by design. Findour HealthAI prioritises patients for review. A qualified reviewer decides what happens next. No clinical action is automated, scheduled or triggered by this software.
🤝 Care Management
Follow-up tracking and review-to-outcome funnel across the reviewed cohort
📝 Executive Summary Report
Auto-generated insights & hospital improvement recommendations
📋 Data Requirements & Metadata
Schema reference, validation rules, and live dataset validation
FINDOUR HEALTHAI · Turning Healthcare Data into Explainable, Actionable Insights FINDOUR SERVICES PRIVATE LIMITED · FINDOUR SERVICES BY BHOPAL SINGH · PROTOTYPE — NOT A MEDICAL DEVICE
🏆 Competition Demo
Step 1 of 8