Predict Customer Behavior From Sales -- 2

I will supply our historical sales data in CSV format, already cleaned of obvious entry errors but otherwise untouched. Your task is to run a full predictive analysis aimed at understanding and anticipating customer behavior—specifically what, when, and how likely each shopper is to purchase again.

Here is what I expect:
• A reproducible model built in Python (pandas / scikit-learn) or R that predicts the next purchase probability and likely product category for each customer.
• Clear feature importance or explanatory output so I know which factors are driving the predictions.
• A concise slide deck or report that translates the results into plain-language insights and recommended actions for marketing and inventory teams.
• All code, notebooks, and any intermediate data files delivered in a shared repository so my in-house analyst can rerun the workflow.
• Performance metrics (AUC, accuracy, or a more suitable measure) on a held-out test set to validate model quality.

I’m happy to answer schema questions quickly so you can move straight into exploratory analysis and modelling. If you prefer additional visualization—Tableau, Power BI, or Matplotlib—feel free to include it.