Multi-cloud Global-Scale Poly-lingual Causal Inference Media Mix Models

Zoom Link--
https://us06web.zoom.us/j/86372329990?pwd=gSDMZ75dwq860nCQbAPSEe9RxuHnUK.1

In this Meetup, John Stanton-Geddes will discuss the core challenges of building media mix models for a major consumer electronics company. Media mix models are experiencing a resurgence of interest as multi-touch attribution approaches are losing relevance in light of online privacy-preserving measures. However, media mix models present a quagmire of challenges that most blogs and vendors gloss over. John will give some real world examples of the challenges my team has faced, and an overview of a path forward using Bayesian methods for Causal Inference. Along the way, he'll describe some of the challenges and joys of working with a globally distributed team in a large company, pitfalls to beware of, and approaches that have helped motivate success.

About The Speaker:
When he's not organizing hockey tournaments, John Stanton-Geddes is a recovering biology PhD who has been working in industry and an member/organizer of the Burlington DS Meetup for 10 years+.