How armaghan6623/no--show-iq- works
NoShowIQ predicts appointment no-show risk from clinic appointment data. The documented training workflow reads the Kaggle dataset, cleans and engineers features, trains and evaluates a classifier, and saves a model artifact; the API loads that artifact and serves single or batch predictions. The API source shown does not call the preprocessing or MongoDB modules, so those connections are not inferred. MongoDB’s prediction logging, training-run recording, and statistics capabilities are visible in its module, but their callers are not shown. The README also describes history and deployment capabilities not substantiated by the sampled runtime source.
Main entry points: run.py Training runner · api.py Prediction API · model.py Train and evaluate · preprocessing.py Feature preprocessing
Components 9
Model training
- Training runner
run.py - Feature preprocessing
src/noshow_iq/preprocessing.py - Train and evaluate
src/noshow_iq/model.py
Prediction API
- Prediction API
src/noshow_iq/api.py
Persistence
- MongoDB operations
src/noshow_iq/mongodb.py
Other components
- Clinic user
- Appointment CSV
data/KaggleV2-May-2016.csv - Saved model artifact
- MongoDB
Connections 10
- Training runner → Appointment CSV: reads
- Training runner → Feature preprocessing: calls preprocessing
- Feature preprocessing → Training runner: returns features
- Training runner → Train and evaluate: trains and evaluates
- Train and evaluate → Training runner: returns model and metrics
- Training runner → Saved model artifact: saves model
- Clinic user → Prediction API: requests prediction
- Prediction API → Saved model artifact: loads model
- Prediction API → Train and evaluate: predicts
- MongoDB operations → MongoDB: reads and writes