Back to Projects
AutomationAnomaly DetectionPythonAlertingBigQuery
Stars Performance Early Warning System
Automated risk detection and alerting for quality measures
December 2023
Data Analyst
Project Overview
An automated system that monitors measure performance trends and triggers alerts when scores drop below critical thresholds or exhibit anomalous behavior.
The Challenge
Star Ratings can fluctuate unexpectedly. Waiting for monthly reports is often too late to intervene. The business needed a way to detect risks immediately.
Key Results
- Reduced time-to-detection for performance drops from 30 days to 1 day.
- Proactively identified 3 simulated measure risks before they impacted end-of-year ratings.
- Streamlined the intervention process for the quality improvement team.
Technical Solution
- 1
Developed BigQuery Scheduled Queries to run daily performance snapshots.
- 2
Wrote a Python script using Z-score analysis to detect statistical anomalies in measure rates.
- 3
Integrated with a simulated email alerting system to notify business owners of risks.
- 4
Created a 'Risk Heatmap' dashboard to visualize areas needing immediate attention.
Tech Stack
AutomationAnomaly DetectionPythonAlertingBigQuery