SyncFit - AI-Powered Caregiver Communication Platform

Inspiration

My huge interest in healthcare and machine learning model that I've been trained for!

What it does

Provides endless health monitoring system for caregivers

Project Overview

SyncFit transforms healthcare monitoring into meaningful family communication. We built an AI-powered platform that monitors elderly patients' health data from wearables, uses machine learning to predict health risks, and generates compassionate, actionable messages for family caregivers through Storyblok's headless CMS.

Why We Built This

• 30% of elderly patients experience preventable health emergencies
• Family caregivers often feel helpless, not knowing when their loved ones need attention
• Current healthcare systems focus on providers, not families
• We needed a solution that transforms cold data into warm, actionable family communication

What Makes SyncFit Unique

Instead of just storing health data, we create content experiences for families:

• AI generates personalized, compassionate messages for caregivers
• Storyblok manages and delivers these messages as rich content
• Families receive actionable guidance, not technical metrics
• The system scales globally through Storyblok's CDN

Storyblok Integration

Content as Communication

We reimagined Storyblok not as a database, but as a caregiver communication platform:

1. Custom Components for Healthcare Content

   • caregiver_alert: Structured health alerts with rich formatting
   • message_content: AI-generated compassionate messages
   • emergency_contacts: Dynamic contact information
   • action_plan: Step-by-step guidance for families
   • resource_links: Educational content for caregivers

2. Content Organization by Severity

   storyblok-space/
   └── caregiver-alerts/
      ├── critical/     # Emergency interventions
      ├── high/         # Urgent care needed
      ├── moderate/     # Monitor closely
      └── low/          # Routine updates
   

3. Real-Time Content Delivery

   • Webhook integration for instant family notifications
   • CDN-powered global content distribution
   • Multi-channel delivery (web, mobile, SMS ready)
   • Visual editor for non-technical staff to customize messages

Key Storyblok Features Utilized

• Headless CMS Architecture: Decoupled content from presentation
• Content Delivery API: Real-time content fetching
• Management API: Programmatic content creation
• Webhooks: Automated family notifications
• Visual Editor: Preview caregiver messages
• Global CDN: Sub-100ms content delivery worldwide

Tech Stack

Backend & AI

• Python 3.9+ - Core application logic
• FastAPI - High-performance async API server
• XGBoost - Machine learning for health risk prediction (95% accuracy)
• OpenAI GPT-3.5 - Generates compassionate caregiver messages
• Pandas/NumPy - Data manipulation and analysis

Content Management

• Storyblok CMS - Headless content management
• Storyblok Python SDK - API integration
• Custom Components - Healthcare-specific content structures

Frontend

• HTML5/CSS3 - Semantic, accessible markup
• JavaScript ES6+ - Interactive functionality
• TailwindCSS - Responsive, compassionate design
• Chart.js - Health metrics visualization

Data Pipeline

• Apache Airflow - Workflow orchestration (optional)
• SQLite - Local data persistence
• JSON - Data interchange format

Code Snippets

AI-Powered Message Generation

# storyblok_caregiver_system/message_generator.py
class CaregiverMessageGenerator:
    def generate_caregiver_message(self, patient_name, severity, 
                                    inactive_hours, activity_decline):
        prompt = f"""
        Create a caring message for family about {patient_name}'s health.
        Severity: {severity}, Inactive: {inactive_hours}h, 
        Activity drop: {activity_decline}%

        Include: empathy, clear explanation, 3-5 actions, reassurance
        """

        response = openai_client.chat.completions.create(
            model="gpt-3.5-turbo",
            messages=[{"role": "user", "content": prompt}]
        )

        return self._structure_for_storyblok(response)

Storyblok Content Creation

# storyblok_caregiver_system/content_manager.py
class CaregiverContentManager:
    def create_caregiver_alert(self, alert_data):
        # Determine folder based on severity
        folder = f"caregiver-alerts/{alert_data['severity']}"

        # Create story in Storyblok
        story = {
            "story": {
                "name": f"Alert for {alert_data['patient_name']}",
                "slug": f"alert-{alert_data['patient_id']}-{timestamp}",
                "parent_id": self._get_folder_id(folder),
                "content": {
                    "component": "caregiver_alert",
                    "patient_name": alert_data['patient_name'],
                    "message": alert_data['ai_message'],
                    "severity": alert_data['severity'],
                    "emergency_contacts": self._get_contacts(severity)
                }
            }
        }

        response = storyblok_client.post("/stories", json=story)
        return response

Machine Learning Risk Prediction

# src/churn_model.py
def train_churn_model():
    # Feature engineering from wearable data
    features = ['steps_mean', 'steps_std', 'heart_rate_mean', 
                'heart_rate_max', 'class_attendance']

    # Train XGBoost model
    model = XGBClassifier(
        n_estimators=100,
        max_depth=4,
        learning_rate=0.1
    )

    model.fit(X_train, y_train)

    # Achieve 95% accuracy
    predictions = model.predict(X_test)
    accuracy = accuracy_score(y_test, predictions)

    return model

Quick Start Demo

Prerequisites

• Python 3.11
• Git
• Storyblok account

1. Clone and Setup

# Clone repository
git clone https://github.com/Rummikub/portfolio.git
cd portfolio/SyncFit

# Install dependencies
pip install -r requirements.txt

# Configure environment (create .env file)
cat > .env << EOF
OPENAI_API_KEY=your_openai_key_here
STORYBLOK_TOKEN=your_storyblok_token_here
STORYBLOK_SPACE_ID=your_space_id_here
EOF

2. Generate ML Predictions

# Create synthetic data and train models
python run_pipeline.py

3. Generate Caregiver Messages

# Generate AI messages for critical patients
python generate_caregiver_messages.py --all-critical-patients

# Or for specific patient
python generate_caregiver_messages.py --patient sarah_wilson --severity critical

4. Start the Caregiver Portal

# Launch API server (KEEP THIS RUNNING!)
python storyblok_caregiver_system/api_server.py

Project Structure

SyncFit/
├── storyblok_caregiver_system/    # Caregiver communication core
│   ├── message_generator.py       # AI message generation
│   ├── content_manager.py         # Storyblok integration
│   └── api_server.py             # FastAPI server
├── src/                          # ML pipeline
│   ├── churn_model.py           # Risk prediction
│   ├── guardian_alerts.py       # Alert generation
│   └── feature_builder.py       # Feature engineering
├── frontend/                     # User interfaces
│   ├── caregiver_portal.html   # Family portal
│   └── dashboard.html           # Admin dashboard
├── data/                        # Data storage
├── models/                      # Trained ML models
├── logs/                        # Documentation
│   ├── DEMO_SCRIPT.md          # Demo walkthrough
│   ├── STORYBLOK_SETUP_GUIDE.md # Setup instructions
│   └── CAREGIVER_FEATURES_SUMMARY.md
└── README.md                    # This file

Architecture

graph LR
    A[Wearable Devices] --> B[Python ML Pipeline]
    B --> C[Risk Prediction]
    C --> D[AI Message Generator]
    D --> E[Storyblok CMS]
    E --> F[Webhooks]
    E --> G[CDN]
    F --> H[Family Notifications]
    G --> I[Caregiver Portal]
    I --> J[Family Members]

Built With

• airflow
• css3
• git
• javascript
• json
• machine-learning
• openai
• python
• sqlite
• storyblok