Customer Churn Prediction — CISC 593 P1

Coursework for CISC 593 (Testing & Verification). The application is a Python machine-learning system that predicts customer churn using UCI Machine Learning Repository data; the grading target is the unit and system testing built around it.

See:

• PropjectProposal.md — original project proposal.
• ArchitectureProposal.md — module layout and mock-first ML strategy.
• TestPlan.md — units under test, methodologies, test matrix, engineer assignments.

Repository Layout

src/churn/         # application code (one subpackage per module)
tests/unit/        # unit-level tests (mirror of src/churn/)
tests/system/      # end-to-end pipeline tests
tests/fixtures/    # committed CSV fixtures used by tests
data/              # raw downloads — gitignored, not required for tests

Setup

Requires Python 3.11 or newer.

python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install --only-binary=:all: -e ".[dev]"

The --only-binary=:all: flag forces wheels and avoids needing a recent C compiler; numpy/pandas builds from source require GCC ≥ 9.3.

Running Tests

# Full suite
pytest

# Just unit tests
pytest tests/unit -v

# With coverage report
pytest --cov=churn --cov-report=term-missing

Engineers

Module	Primary author	Reviewer
preprocessing	Emmanuel Ndone Suum	Mirriam Chemutai Ronoh
evaluation	Janus Thor Kristjansson	Zeyu Wang