How erickyegon/insurancepremiummodel works
This is a health-insurance premium prediction application with an ML training lifecycle, prediction API, monitoring views, and retraining support. The principal user flow is request data through the FastAPI app, prediction pipeline, and model deployment component to a premium result. Training is separately orchestrated through ingestion, validation, transformation, model training, and evaluation. Monitoring and retraining capabilities are documented and partly visible in the app excerpts; their deeper integration is uncertain. The graph uses the supplied source excerpts and tree, not unsampled implementations.
Main entry points: app.py FastAPI App · training_pipeline.py Training Pipeline · prediction_pipeline.py Prediction Pipeline · model_deployment.py Model Deployment
Components 19
Prediction Serving
- FastAPI App
app.py - Prediction Pipeline
InsurancePremiumPrediction/pipeline/prediction_pipeline.py - Model Deployment
InsurancePremiumPrediction/components/model_deployment.py
Model Training
- Training Pipeline
InsurancePremiumPrediction/pipeline/training_pipeline.py - Configuration
InsurancePremiumPrediction/config/configuration.py - Data Ingestion
InsurancePremiumPrediction/components/data_ingestion.py - Data Validation
InsurancePremiumPrediction/components/data_validation.py - Data Transformation
InsurancePremiumPrediction/components/data_transformation.py - Model Trainer
InsurancePremiumPrediction/components/model_trainer.py - Model Evaluation
InsurancePremiumPrediction/components/model_evaluation.py
Monitoring Retraining
- Monitoring Metrics
app.py - Drift Detection
mlops/monitoring/data_drift.py - Alerting
mlops/monitoring/alerting.py - Retraining Task
app.py
Deployment Integration
- Azure Deployment
deployment/azure/azure_deploy.py - Azure Scoring
deployment/azure/score.py
Other components
- Business User
- Premium Result
- Source Data
Connections 20
- Business User → FastAPI App: submits request
- FastAPI App → Prediction Pipeline: calls prediction
- Prediction Pipeline → Model Deployment: loads and predicts
- Model Deployment → Premium Result: returns premium
- FastAPI App → Premium Result: returns response
- Training Pipeline → Configuration: gets settings
- Training Pipeline → Data Ingestion: starts ingestion
- Source Data → Data Ingestion: provides data
- Training Pipeline → Data Validation: validates splits
- Training Pipeline → Data Transformation: transforms splits
- Training Pipeline → Model Trainer: trains model
- Training Pipeline → Model Evaluation: evaluates model
- FastAPI App → Monitoring Metrics: builds dashboard metrics
- FastAPI App → Drift Detection: drift route integration unclear
- FastAPI App → Alerting: alert integration unclear
- FastAPI App → Retraining Task: queues background task
- Retraining Task → Source Data: ingests retraining data
- Retraining Task → Data Validation: validates retraining data
- Azure Deployment → Azure Scoring: deploys scoring
- Azure Scoring → Prediction Pipeline: serves predictions