How idharshan/ml-based-vehicle-predictive-maintenance-system-with-real-time-visualization works
This is a vehicle engine-condition prediction demo: a user enters sensor values in a Streamlit interface, requests a prediction, and sees a normal/warning result with a confidence value. The README describes a broader predictive-maintenance workflow, including preprocessing and GBM training, but those notebook details were not sampled; the app itself uses manual sliders rather than a live sensor feed. There is also a filename mismatch: app.py loads model.pkl, while the tree contains hhmodel.pkl, so their connection is unverified. Maintenance-date estimates and scheduling are documented but not implemented in the sampled app.
Main entry points: app.py Sensor input UI · engine_data.csv Engine dataset · data_preprocessing.ipynb Data preprocessing
Components 7
Prediction interface
- Sensor input UI
app.py - Condition and confidence
app.py
Model workflow
- Condition inference
app.py - Data preprocessing
data_preprocessing.ipynb - Model artifact
hhmodel.pkl
Training data
- Engine dataset
engine_data.csv
Other components
- Fleet user
Connections 4
- Fleet user → Sensor input UI: enters sensor values
- Sensor input UI → Condition inference: requests prediction
- Condition inference → Condition and confidence: returns class and confidence
- Engine dataset → Data preprocessing: training data