|
MickeyFinn posted:There are 4 hours of lectures a week and the rest of the time is spent working on projects and doing interviewing/networking. They give you a project every week, tell you a little about it and then leave you to figure it out on your own. The interviewer framed it as a way to build your confidence about learning to tackle problems. He also specifically told me we'd be using Google and reading a lot of documentation. We were strongly encouraged to attend in person because getting help and collaboration is easier in person (note in-person is $17k, remote is $9k). He also said there is a Slack channel for the class in-person and remote alike. That is such a weird "interview". I just finished Metis, a different bootcamp, in December, and I'm starting a really nice job as a data scientist in two weeks. I also got into the Data Incubator at the paying "Scholar" levels, and I declined because they were a little skittish about the costs and specifics, and I generally got a weird vibe from them. The classroom instruction particularly seems light. At a certain point, the best way to learn data science is to do data science, but Metis had at least two hours of hands-on instruction per day, and frequently more. Metis is also 12 weeks instead of 8, and is comparable in costs. There is an element of selling shovels to the miners about the bootcamps, but Metis was upfront about the costs, and I thought the quality of instruction was on par with, and frequently better than, the traditional schools I attended previously. I'd say my cohort was solid, and they weren't all handpicked geniuses. In my opinion, the most important thing in data science is figuring out how to translate whatever messy human problem you're dealing with into something with labels and features that show some kind of pattern. Once you're at the point where you're running classification algorithms on tidy dataframes, some kind of results are guaranteed. There are always improvements to be made, and that's the difference between craftsmen and people who are hammering away with bits and bobs of code from Stack Overflow, but the first and most important step is framing the problem correctly. I came in with a social sciences PhD, which was really good preparation for figuring out how to ask interesting questions. Metis did not do a great job explaining how to do that in 12 weeks, but on the other hand, it took me five years in a PhD program to figure out research design, so if they were to manage teaching that way of thinking quickly, they'd be downright uncanny.
|
#
¿
Feb 15, 2019 06:12
|
|
|
|
| # ¿ May 17, 2024 16:18 |
|