Tech IndustryJan 6, 2019
Intuitqbse

Product analyst vs data scientist at uber

From my understating one role is extremely technical and the other is extremely non technical. Is that true?

Add a comment
Uber Deer Jan 6, 2019

Not really. Its just level of knowledge. Product analysts just usually have less experience and statistical expertise.

PayPal Guangzhou Jan 6, 2019

My understanding is that as a product analyst you will do more of AB testing, general reporting, funnel optimization etc but a data scientist focuses on a particular problem and uses machine learning to solve it. Make sense?

Amazon babycorgz Jan 6, 2019

dude I'm from Guangzhou too

New
zorkan Jan 6, 2019

Oh shit you guys know each other?

New
zorkan Jan 6, 2019

Data scientists design/PoC (proof of concept) new products that leverage data the company has collected (usually through ML or statistical models) -they tend to hold a PhD in one of Math/Stats/EE/CS. Analyst roles are more concerned with reporting, generating internal business insights, ect.. For the most part, data science is an R&D role. And analyst is a somewhat technical business/marketting role. Ymmv at different companies though. I've seen non-tech companies title their SWEs as "Tech Analysts", or slap "Data Scientist" onto any role that touches data.

Quora iSjP52 Jan 8, 2019

I think the op gets the general distinction, they want a specific answer about how that general distinction applies to a specific company’s practices

New
muyH32 Jan 24, 2019

(I don't work at Uber, but know many product analysts and data scientists who do.) Your understanding is correct. The data scientists tend to have PhDs in technical fields (stats, operations research backgrounds), the product analysts tend to be Bachelor's Degrees in Econ or non-technical fields, often people who used to be marketing analysts, business analysts, etc. I've also heard from friends that even the product analysts managers/directors are extremely non-technical; for example, the directors/leads of that group don't know what a t-test or p-value is. Edit to add: I've talked to a few ML engineers and managers at Uber, and all of them seem to find the data scientists useless. They all said the data scientists were doing R&D things that didn't scale, were academic solutions to extremely specific problems, and engineers had to reinvent and reimplement anything as a result. May be a small sample, though.