🌍 Overview
This project builds an AI-driven geographic expansion model for Prescribed Pediatric Extended Care (PPEC) of Palm Beach, which had reached capacity. Using public health and socioeconomic data, I designed a zip-code-level clustering framework to identify underserved families, forecast demand, and guide resource allocation for ethical, data-backed growth.
📘 View Heatmap + Case Study on GitHub →
PPEC centers faced the challenge of expanding without compromising care quality. Leadership needed a scalable, data-backed approach to answer three critical questions:
Where are the highest-need pediatric populations? Which areas should be prioritized for outreach and expansion? How can resources (staff, transport, funding) be allocated most effectively?
Gathered multi-source public data at the zip-code level, including:
Child population density Household income levels Medicaid percentage and insurance coverage Distance to existing PPEC centers
Applied K-Means clustering to segment Florida zip codes by opportunity. Categorized regions into 🟢 High Potential, 🟠 Moderate Potential, and 🔴 Low/Saturated zones. Used statistical weighting to combine socioeconomic and accessibility variables for better prioritization.
Built an interactive Folium heatmap displaying geographic clusters and service gaps. Added color-coded overlays for clear executive communication. Automated data refresh through AWS Lambda and Amazon S3 for recurring updates.
Generated five actionable insights that informed expansion strategy:
Zip codes 33413, 33415, and 33417 show high need but low service availability. “White space” zones have high Medicaid coverage yet remain underserved. AI clustering supports franchise targeting, not just new center builds. Zip-level precision enables hyperlocal marketing with better ROI. Resource load optimization ensures balanced outreach, staff, and transportation planning.
Developed two implementation tracks for leadership execution: Resource Allocation Plan: Grouped zip codes into High, Moderate, and Low priority clusters for staffing and outreach. White Space Zone Playbook: Identified untapped regions ideal for mobile PPEC units, school tie-ups, and community program expansion.
AWS Services:
Amazon S3
AWS Lambda
Amazon SageMaker
Amazon QuickSight
AWS Glue
Amazon EC2
Python
Folium
Scikit-Learn
US Census Data + Florida AHCA APIs
AI Clustering & Predictive Modeling
Geo-Spatial Data Visualization
Public Health Analytics
Strategic Growth Planning
Key Improvements Achieved:
✅ Community Coverage: Expanded from 3 centers to 15 high-impact zones
✅ Outreach Efficiency: Increased targeting precision by 40%
✅ Resource Utilization: Shifted from manual allocation to fully data-driven prioritization
✅ ROI on Expansion: Projected 2.3× return on the pilot phase
This project enabled PPEC of Palm Beach to:
Expand strategically into underserved pediatric communities Use AI-backed evidence to justify grants, funding, and licensing applications Plan staffing and outreach with measurable social impact Establish a repeatable, scalable data model for future state-wide PPEC expansion