How saiteja74/diabetes_prediction works
This repository is a Streamlit app for exploring an early-stage diabetes dataset and estimating diabetes risk from user-entered symptoms. Users choose the EDA or ML section: EDA loads supplied CSVs for summaries and plots, while ML collects and encodes patient attributes before displaying a model prediction and probability scores. The graph reflects the shown routing and data use. One uncertainty: ml_app.py requests a model filename absent from the tree, which instead contains decision_tree_model.pkl; their relationship is therefore not asserted.
Main entry points: eda_app.py EDA interface · app.py Menu and routing · ml_app.py Risk questionnaire · diabetes_data_upload.csv Diabetes records
Components 11
Streamlit App
- Menu and routing
app.py
Data Exploration
- EDA interface
eda_app.py - Diabetes records
diabetes_data_upload.csv - Cleaned records
diabetes_data_upload_clean.csv - Age frequencies
freqdist_of_age_data.csv - Data summaries
eda_app.py - Exploration plots
eda_app.py
Risk Prediction
- Risk questionnaire
ml_app.py - Feature encoding
ml_app.py - Risk prediction
ml_app.py
Other components
- App user
Connections 10
- App user → Menu and routing: selects section
app.py - Menu and routing → EDA interface: routes to EDA
app.py - Menu and routing → Risk questionnaire: routes to ML
app.py - EDA interface → Diabetes records: loads records
eda_app.py - EDA interface → Cleaned records: loads cleaned data
eda_app.py - EDA interface → Age frequencies: loads frequencies
eda_app.py - EDA interface → Data summaries: renders summaries
eda_app.py - EDA interface → Exploration plots: renders plots
eda_app.py - Risk questionnaire → Feature encoding: encodes answers
ml_app.py - Feature encoding → Risk prediction: predicts risk
ml_app.py