3 ms·
I think the real trick is to never use Jupyter / notebooks to begin with.
by dmicah 5y ago
I think the real trick is to never use Jupyter / notebooks to begin with.
- idomi 5y agoNot really, data scientists love notebooks, it's pretty inefficient to move out of the notebooks.
- FridgeSeal 5y agoThen maybe as data scientists we should stop writing production code in notebooks and instead write and deploy it properly, like the rest of the development community does, and stop papering over laziness and bad habits. It’s only inefficient to move out of notebooks if you’ve written poorly structured, messy code in the first place. We wouldn’t accept that out of other devs, so I’m not sure why data scientists should magically get a pass.
- paulgb 5y agoI think seeing this as data scientists “getting a pass” is looking at it through the wrong lens. Data scientists and quants tend to be good at statistics, data modeling, and solving business problems, not software engineering except as a means to an end. You might be able to find a few people who have a combination of those skills, but you’re picking from a very small pool.
- FridgeSeal 5y agoWasn’t the point of the data scientist/ML engineer that they could combine the software engineering skills with the statistical and knowledge know-how? Which is why they’re paid so much? If they can’t do the former, may as well just hire a software developer and a statistician for near-cheaper and have them work together - you save money and get better code.