AI Predictive Modeling for Structured Data

I have several well-organised spreadsheets and database tables that hold historical business figures. The task is to turn this structured data into practical, production-ready predictive models that I can use for forecasting and decision support.

You will handle the full analytical workflow—exploring the datasets, selecting and engineering features, training the models, and validating their performance. I expect clear, reproducible code (Python with pandas, scikit-learn or a comparable AI/ML stack fits best), accompanied by a concise report that explains the modelling approach, key findings, and how to apply the final model(s) to new data.

Deliverables
• Clean, commented source code and notebooks
• Trained model files and any preprocessing pipelines
• A brief performance report including chosen metrics and validation results

Acceptance criteria
• Model meets an agreed accuracy benchmark on a held-out test set
• All steps are reproducible on my machine with provided instructions
• Documentation is clear enough for a non-technical stakeholder to understand the insights and use the predictions.