JunDG92/property-maintenance-priority-baseline

Property Maintenance Priority Baseline

This repository contains a small baseline model artifact created with property-ops-ml, a reusable Python toolkit for property-management analytics workflows.

Model Summary

The model is a transparent logistic-regression baseline for maintenance-priority review. It scores work-order-like records and returns a review-support label.

This is not a universal production model. It is a reproducible example showing how property-management teams can package a model artifact, document it, and load it for inference.

Intended Use

Appropriate uses:

• workflow demonstration
• reproducible ML packaging
• review-support prototyping
• data-cleaning and feature-engineering practice
• baseline modeling before retraining on approved internal data

Out-of-scope uses:

• automated maintenance decisions
• tenant/resident decisions
• production dispatching without validation
• use as a substitute for property-management judgment

Training Data

The included artifact was trained on a tiny example file bundled with the property-ops-ml GitHub repo. It is intentionally small and exists only to demonstrate the workflow.

Teams should retrain and validate the model using approved internal work-order labels before operational use.

Files

• sample_maintenance_model.json: portable model artifact
• inference_example.py: first-use inference script
• sample_record.json: sample work-order-like record
• requirements.txt: minimal runtime dependencies

First Use

pip install -r requirements.txt
python inference_example.py

Expected output:

{
  "score": 0.8,
  "label": "review",
  "threshold": 0.5
}

Exact score may vary if the artifact is regenerated.

Responsible Use

This model is for review support and learning. Any real deployment should include data governance, validation on approved local data, threshold calibration, drift monitoring, and human review.