Background: Online storefronts advertise and promote select products on their home page, such as Amazon or Temu in the case of e-commerce, or Instacart or Uber Eats in the case of food delivery. However, in situations where shopping platforms connect buyers to sellers but do not control inventory themselves, at any given moment, 10-20% of advertised items are actually out of stock.

Objective: Predict which products are likely to be out of stock, so they can be removed from the advertising list.

Alternatively, predict metrics on customer behavior (recommendation systems, next purchase date, or next purchase product type).

Data: Many scraped datasets are available online, such as
* https://www.kaggle.com/datasets/thedevastator/unlock-profits-with-e-commerce-sales-data/data?select=Amazon+Sale+Report.csv
* https://www.kaggle.com/competitions/grupo-bimbo-inventory-demand/data