Hi folks, I have been a data scientist ( analytics) for quite some time and I'm finding it that in far too many companies data scientist is a glorified SQL monkey. What that means is that their work involves working closely with the stakeholder, doing sizing and estimation for opportunity, and poorly designed a/b tests, and only that. The roles do not require statistical knowledge. Upper management in data science in these roles neither has statistical knowledge nor do they encourage it. Which companies have a truly data driven culture? And by data driven I don't mean that they need to bring data to build a case, because often you can twist the data to tell the story that you want to. But the companies where analytics means having a strong understanding of statistics, and a healthy skepticism of the data. Where people define a hypothesis, then gather data and analyze it and accept it without tinkering further to twist it to tell the story that they wanted to tell. Edit: or should I just change tracks to ML to scratch that technical itch?
Worked for Samsclub a few years ago. I was surrounded by phd grads in stats and some people even from bio + stats background working in retail. They were working on proper end to end hypotheses without looking at the data. Used the data to test the hypotheses instead of the other way around. Also, worked on optimization problem statements in a supply chain setup. Sam's club as a website grew a lot in the last 2-3 years. Probably, that's why. One gotta be in the right team which gets really tricky in bigger orgs.
How are things at Spotify?
Kinda growing on me. It's been ok. WLB >> & Work from Home - Permanent (my role). Cantt complain a lot tbh
Google's DS track
Does that involve ML? Or more focus on experimentation?
This is the best answer here. It involves rigorous stats at a minimum. Depending on the team it can be ML or Experimentation or many other things.
I have experienced this same thing. Looking forward to seeing others responses
Stitchfix’s DS team is well respected
There’s some director and above level turmoil but the work is legit. That’s the first time an interview made me sweat
Sorry but... What.
Data science should involve ML
Says who? It's a super broad field
@linkedin you're not a scientist if you're a SQL monkey. It's not Data Science without ML.
Don’t join GS unless you want to learn a proprietary language (Slang) that isn’t necessarily better than Python and you can’t use anywhere else.
We’re doing some fun stuff at Morning Consult - data intelligence via polling survey data.
How is DS at credit karma . Is it glorified sql monkey
DS at credit karma is ML focused. Analytics is more SQL monkey.
Startups are where you have full control over the techniques you wanna be using. If they have a DS team, very likely they’re gonna wanna listen to whatever you have to say. And that would also mean more opportunity for you to try new statistical techniques to create value or telling stories.
Yeah, but with startups, they usually don’t have the infrastructure, or even the data, to do any real analyses… with startups you’ll be doing a whole lot more data engineering than data science…
Yep, can second this. I was brought in for “research” and all I do is shitty data eng tasks
Academia. And even then people are p-hacking their way to novel results.
Its not worth it. Too much effort and politics for low tc. Post tenure life can be great though
The path is to do 10-15 years at a high paying job and then to semi retire to teach at a community college. No pressure. You get summer off if you want it. You can travel. You can take classes on the side.