HealthPulse AI

Hospitals upload their historical patient admission data along with real-time pollution, weather, festival calendar, and epidemic alerts into the system. Tools like **Rows and Firebase **store and organize this data for analysis.

The system merges all datasets and cleans the data to create a **structured input **for AI analysis. This ensures that all factors like AQI, festivals, epidemics, and past patient trends are accurately represented.

Using Flowise AI with RAG and fine-tuned LLMs, the system analyzes this combined data to predict potential patient surges in hospitals, taking into account environmental conditions, local festivals, and ongoing epidemics.

The AI also references hospital guideline PDFs and operational protocols using RAG and Ollama (offline and private) to generate specific actionable recommendations. These include the number of additional staff required, extra beds, medicine stock updates, and patient advisory messages.

If the predicted patient count exceeds safe thresholds, the system automatically triggers real-time alerts **to hospital administrators via Twilio, Email, or Webhook nodes. The predicted surges and recommended actions are displayed on a live dashboard built with frontend, showing trends, real-time AQI, festival impacts, and epidemic status. All predictions, advisories, and alerts are **logged into Firebase or Rows **for historical tracking, monitoring, and future AI model training. After the actual events, the system compares predicted vs. actual patient inflow **and uses this data to fine-tune AI models, improving the accuracy of future predictions. Optionally, patients in the affected region can receive advisory notifications about preventive measures, helping hospitals manage patient inflow and reduce overcrowding. and cleans the data to create a structured input for AI analysis. This ensures that all factors like AQI, festivals, epidemics, and past patient trends are accurately represented. Data Collection Hospitals upload historical patient data.

Live data pulled from AQI & weather APIs, festival calendar, and epidemic alerts.

Tools: Rows, Firebase, OpenWeatherMap API, AQICN API, Calendarific API

Data Integration & Cleaning Merge and structure all datasets for AI processing.

Tools: Rows, Firebase, Python/Node.js

AI Prediction Engine Predict patient surges considering pollution, festivals, epidemics, and historical trends.

Tools: Flowise AI, RAG, Fine-tuned LLMs

Guideline-Based Advisory Generation Reference hospital protocols and guidelines to suggest staff, beds, medicine, and patient advisories.

Tools: RAG, Ollama (offline & private)

Threshold Detection & Alerts If predicted patients exceed safe limits, trigger real-time alerts.

Tools: Twilio, Email, Flowise Webhook Node

Dashboard & Visualization Display predictions, advisories, AQI, festival, and epidemic trends on live dashboard. Logging & Monitoring Store all predictions, alerts, and AI actions for historical tracking and analysis.

Tools: Firebase, Rows, Google Sheets

Continuous Learning & Feedback Update AI models based on actual patient inflow to improve future predictions.

Tools: Flowise AI, Fine-tuned LLMs

Optional Patient Advisory Send preventive advice to patients to reduce overcrowding.

Tools: Twilio, Email, WhatsApp API