Predict Purchase Behavior from Customers

I have a structured customer-level dataset and the central aim is to turn it into a reliable predictor of future purchase behavior. The work revolves entirely around outcome prediction, not trend-spotting or process optimization, so every step—from cleaning the raw customer records to validating the finished model—should directly support that goal.

Here is the flow I envision:
• Data preparation: handle missing values, engineer features that truly influence buying decisions, and document every transformation in clear, reproducible code (Python, R, or SQL—whichever you prefer).
• Model development: test several supervised learning techniques, benchmark them with appropriate metrics (AUC, precision-recall, or another mutually agreed KPI), and converge on the best performer.
• Insight & deployment assets: deliver a concise report explaining key drivers of purchase likelihood, the trained model files, and a repeatable script/notebook so I can refresh predictions whenever new customer data arrives.

Acceptance criteria
• End-to-end script runs without manual tweaks on my sample data.
• Minimum predictive lift over baseline (to be finalized together).
• Clear explanation of how the model can be integrated into my existing workflow or dashboards.

If you are comfortable building purchase-propensity models and presenting the results in plain language as well as code, this should be a straightforward engagement.