Related Posts
Very proud of her toy she brought to me 😂
Hi all,
I need 11 ♡ for my dm to be enabled.
Thanks 🙂
What are your summer plans?
Additional Posts in Consulting
My weight positively correlates with my utilization.
I am a senior with 40% utilization.AMA
What do y’all like to do on the weekends?
Has anyone else's MQMs not posted yet?
MicroStrategy or Tableau?
New to Fishbowl?
unlock all discussions on Fishbowl.
Naw software engineering. They’re real coders. Data scientists are usually not professional coders. That said, it’s a different skill set and probably easier to get into if you weren’t a CS major.
As a data scientist the programming you are doing is probably hacking together Python or R scripts built from scratch by yourself. You use them once and throw away.
As a software engineer you try to fix and update an incomprehensible mess of legacy code written by 100 developers before you and there’s none of them left to answer to your questions about that the hell the code is supposed to do.
That’s unless you are joining a startup building a new product but frankly you’re not there yet.
There are all sorts of data scientists and software engineers so it is a little hard to generalize.
Are you trying to create the next xgboost? That's hard work. Probably harder than most software engineering.
Are you a data scientist performing one off analyses using of the shelf packages? You need good statistical knowledge but the application isn't to hard.
Are you a data scientist performing analysis and writing code that will be integrated with business critical systems that need to be performant? That's hard.
Most data scientists start as the middle kind.
Data science. It’s programming + stats + business
Bcg2 yea that’s what I’ve read too. They dabble in python but like 60% of their time is spent cleaning data
Lol data science is mostly doing the statistics and modeling, not to look down on it, but just that you seem to not have any idea what it entails.... Machine learning engineer is the middle man of those two but you're not creating the models, just implementing them
Software engineering. I'm a data scientist.
That said data scientists building production coflde probably have it worst.
Director 1 with the little I know I agree with you, but curious as to why you say this? Also what do you mean by production cofdle?
So as a data scientist here, I will tell you we do the 3rd kind
Broad strokes not judgements. I switch between two and three depending on project needs.
Would you guys say that the nature of the work can be ambiguous? Reason I ask is because with all the projects I’m on in technology the work is feel like with data science maybe it’s not the case. Maybe difficult and that’s ok as long as it’s more black and white
Last post was grammatically fucked I’m drunk sorry
Not sure what you mean, but you'll know if you are doing data science.
D1 I meant is the nature of data science project work ambiguous? Or is it pretty clear cut on what needs to be accomplished?
Like all things it depends. Things that seem clear cut aren't necessarily clear cut. For example I spent quite a bit of time on a working capital forecasting model for a large $200B+ rev per year org. The ask is pretty clear in that we need to improve the finance orgs working capital forecast.
The ambiguity lies in how to accomplish it. Are we creating top down, bottom up, or middle out forecasts? What is the right level of forecast granularity? How do we handle forecasts for new products? How do we integrate the forecasts with management systems? How do we integrate user feedback into the forecasts? To what extent do we need to be able to explain the models? What technology or data constraints do we need to work with?
I could go on but most organizations haven't thought through most of these problems, and wouldn't think to do so, so it isn't just the model building but also figuring out what works in the org.
I've also worked on a number of projects where we are just fishing for insight. Those can be frustrating.
Gotcha, thanks