What statistical techniques do you use in Facebook’s marketing science division? Is it anything beyond regression? What about Monte Carlo or something similarly advanced? And is it mainly using Excel or does anyone use R?
SAS????
Wow
What about in terms of statistical techniques? Not sure what else is needed other than regression when you’ve only got Facebook ad metrics.
Probably need to know some Excel but I'd think R is more important. They're going to focus on sampling and experimental design, interesting metrics you come up with on your own and ways to visualize them. It should be intense but basic knowledge. I can't see them asking about complex viz etc. Be familiar with hypothesis testing even though it isn't necessarily relevant (since n gets large, all differences are significant). Also understand repeated measures and bonferroni correction.
SQL is important too
Important just to pull the data to csv? I thought all the data needed was available to export from the Facebook ads GUI? Or do you need to pull your data from a database?
I didn't work in this area when I was at FB but I'd imagine you'd be pulling data to a CSV if you use SQL. Even at the SWE level the queries asked in the interview are basic. Select, group by, order by and a join. I imagine the more you can do in that area the higher the level you'd be hired at. Most of the data probably comes from UI but I wouldn't put all my eggs in that basket just in case
I heard the use SAS and sql mainly. (replying to your third question only)