Stan applications in business/marketing analytics


#1

Hi everyone,

Hope this message finds you all well.

I’m looking for STAN applications in business analytics. So far I have browsed through GitHub repositories but have not encountered any a meaningful hands-on examples solved by STAN. I’d be glad if someone could provide me a few links with relevant case studies.

Also, I’m trying to captivate the interest of my peers at work, but I find it difficult to convey these advanced statistical concepts to non technical audiences. Whenever I touch upon the subject “Bayesian analysis”, most of my colleagues think it’s a kind of a “voodoo magic box”. Exciting but with no relevant business value…
Any hints/books that can help me out on this?

Many thanks,


#2

Check out this bit from the Savage: https://modernstatisticalworkflow.blogspot.com/2017/03/aggregate-random-coefficients-logita.html

This one too: https://rpubs.com/jimsavage/ag_rcl_talk


#3

I’m also really interested in this. I’m trying to use Stan to simulate and model revenue, and can’t find any similar examples online.


#4

Facebook has a time series forecasting tool available in R and Python called prophet which is built on top of Stan. I first heard about it while watching one of the StanCon 2018 presentations (starts about 18 minutes in). I think the corresponding case study is here

Also, it’s not STAN, just Stan :)


#5

It is usually a better idea to model quantity sold as a function of price, etc. (classic elasticity setup) and compute the revenue post hoc.


#6

Hi Eric,

Many thanks for your input.
I’m just going through the links you posted.

I reckon Bayesian analysis are of great use to online retailers, and ultimately applicable to dynamic pricing algorithms.

Have you ever seen Bayesian analysis applied to price elasticity problems or dynamic pricing algorithms?

Cheers


#7

Dynamic pricing is a loaded term, so not sure.

As to elasticity of demand for the purposes of pricing, yes. For example, if you go to Amazon and

order a backlist (1y+) e-book that was published by Penguin Random House, there is a

good chance that we (Stan) priced it.


#8

Alex Braylan and Daniel Marthaler presented their work on estimating elasticity and dynamic pricing during holidays at StanCon Helsinki. Youtube isn’t loading the video right now, but I was able to find the notebook here.