⌚ Real-Time Health Monitoring & Predictive Analytics
Predictive analytics system leveraging wearable devices and real-time cloud infrastructure to detect heart
disease and deliver personalized health insights.
📌 Project Background
To enable scalable, secure, and real-time health analytics, I designed and deployed a full-stack
monitoring and visualization system using Google Cloud:
- Built a full-stack dashboard using GCP (Cloud SQL, Functions) and ECharts to monitor system and
health metrics in real time.
- Modeled a multi-level asset structure (site > floor > room > rack), simulating a DCIM-style
deployment framework.
- Implemented user access control and role-based permissions across dashboard components.
- Gained hands-on experience with DCIM concepts such as structured asset tracking, data consistency,
and live dashboard monitoring.
🔧 Tech Stack
Python
Flask
BigQuery
Google Cloud Platform
Gaussian Naive Bayes
SVM
Logistic Regression
ECharts
ETL
🚀 Core Contributions
- Developed a heart disease prediction model using Gaussian Naive Bayes (GNB) with 85% accuracy based
on real-time sensor input.
- Processed over 50GB of raw health data with Python, and implemented an ETL pipeline using BigQuery
for scalable streaming.
- Improved accuracy by 10% through model comparison (GNB, SVM, Logistic Regression) and hyperparameter
tuning.
- Built a real-time dashboard with Flask + ECharts, optimized BigQuery streaming to reduce latency by
20%.
📊 Results Overview
| Category |
Outcome |
| Prediction Accuracy |
85% (Gaussian Naive Bayes) |
| Improvement from Tuning |
+10% via SVM, Logistic Regression |
| Latency Optimization |
-20% with BigQuery streaming |
| Data Volume Processed |
50GB+ wearable device data |