Healthcare Analytics / Data Engineering
Payer Claims & Cost Analytics Platform
A payer analytics solution designed to evaluate claims, eligibility, utilization, PMPM cost, population trends and healthcare outcomes across member populations.
The Business Problem
Payer organizations need a reliable way to connect claims, eligibility and utilization to understand what is driving cost across member populations.
The Vision
A trusted claims and cost model where PMPM, utilization and population trends reconcile and can be explored by plan, population and service category.
My Role
Challenges
The Solution
Detailed solution narrative in development.
Inside the Solution
Behind the Build
A polished interface is only the visible layer. This section shows the data, architecture, engineering and design decisions behind the finished solution.
-- Synthetic data example
SELECT
e.plan_id,
DATE_TRUNC('month', e.month) AS month,
COUNT(DISTINCT e.member_id) AS member_months,
SUM(c.paid_amount) AS total_paid,
SUM(c.paid_amount)
/ NULLIF(COUNT(DISTINCT e.member_id), 0) AS pmpm
FROM eligibility_months e
LEFT JOIN claims c
ON c.member_id = e.member_id
AND DATE_TRUNC('month', c.service_date) = e.month
GROUP BY 1, 2;sql
Member months & PMPM
PMPM is only meaningful when the denominator is right. Member months come from eligibility — not claims — so members with no utilization still count toward the population cost.
Solution Architecture
- 1Source Systems
- 2Data Ingestion
- 3SQL / Python / ETL
- 4Databricks / Microsoft Fabric
- 5Data Model / Semantic Layer
- 6Power BI / Application
- 7Leadership / Operations / Clinical Users
Technology Stack
Data Engineering
- Databricks
- ETL
- Data Modeling
Programming
- SQL
- Python
- Pandas
Business Intelligence
- Power BI
Security, Privacy & Governance
Healthcare and AI solutions require more than functionality. They require deliberate attention to privacy, security, access, governance and responsible use.
Designed using HIPAA-aligned privacy and security principles.
Testing & Data Quality
- Data reconciliation
- Source-to-target validation
- Duplicate detection
- Row-count validation
Outcome
Demonstrated / Simulated Outcome
Outcome documentation is in progress. Results will be published here once validated against the demonstration dataset.
What I Would Build Next
- High-cost claimant forecasting
- Automated data freshness monitoring
Related Projects
Related skillsHealthcare Analytics / Business Intelligence
LTSS & Dual Eligible Analytics Command Center
A healthcare analytics environment designed to help leadership monitor LTSS and dual-eligible populations across enrollment, eligibility, utilization, case management, quality and cost.
Privacy, Security & Analytics
HIPAA Privacy & Security Incident Intelligence Platform
A centralized analytics and incident-management solution designed to transform fragmented privacy and security data into actionable risk intelligence for healthcare leadership.
AI & Analytics
Healthcare AI Analytics Copilot
An AI-assisted analytics solution designed to help healthcare leaders ask business questions, explore governed data and receive clear analytical explanations supported by underlying data.