Great slide deck, thanks!
I am aware of papers where the graph is assumed to be known (e.g. arxiv.org/abs/2103.06392), and indeed I can construct one from the data that I have. But my concern is whether using the empirical graph ignores selection bias from customers choosing to form edges with certain products.
For outcomes, think of long-term customer spend, retail membership signup, or something else that is affected by the customer experience (the aggregate experience from products they interact with) but are not immediate consequences of a single product interaction.
Suppose an A/B test on a retail platform where products are randomly assigned into treatment or control, and outcomes are measured for customers. Edges are customer-product interactions (e.g. viewing a product). The treatment affects the desirability of the product, the likelihood it surfaces, etc.
Do any #stats#econometrics#causal experts have recommendations for studying bipartite experiments with endogenous edges between the diversion units and the outcome units?
itās a weird partition of services: thereās a job board for serious folks and then a feed of cringey wannabe influencers peddling nonsense, phony inspirational stories, and infographics that remind you the difference between a left and right join as if thatās so freaking complicated
when you log back on to your work computer after a long weekend and five million windows/tabs open back up
does this only work in december or can i put santa in a speedo out in july to signal my interest in having a neighborhood pool party
maybe exposure to the former drove demand for the latter, thatās how it works in my household