Some general questions

Hi community,

I am still working on the courses and I might have missed something.

  1. Do we have courses about how to prepare data ? I know in our homework assignment, the data are usually well prepared and we can call a function, then the right format data is loaded in my notebook, but in the real world, that won’t be the case. How do we know what format is the right one for my problem ?

  2. In real world, how does a jupyter notebook fit in the development process ? I am a regular software engineer, can’t imaging how to “deploy” a notebook in production. Do we use something different than that to “hold” our model ?

  3. Do we have short courses about using those graph plotting tools ? I feel they are pretty useful in working on assignments.

@Shuwen1 so I can only suggest from my personal experience which I think is different from other learners.

  1. This course will tell you how to muck about with data-- but the class is taught in R, not Python. Otherwise you can look up anything to deal with ‘wrangling’ to find more resources. (https://www.edx.org/learn/data-science/harvard-university-data-science-wrangling)

  2. I could be wrong here, but I believe Jupyter Notebook is more a ‘show and tell’ sort of platform. I can’t imagine anyone deploys these things in practice. But you always have Python on the backend (of course then you probably need to develop some sort of ‘frontend’. [Perhaps Django, Flask, or similar])

  3. I also don’t know about that either (as regards to what is offered here). I mean Matplot lib is fairly standard, but if you want to do some more fancy stuff learn/take a class on Seaborn (at least as applies to Python).

Thanks for the input @Nevermnd ,

  1. I guess I might also need to learn R :smile:.
  2. So…the model itself will probably be some python script version of the notebook then…
  3. :+1:
1 Like

Being ‘multi-lingual’ never hurts-- And if you really want to have some fun (which I don’t know yet) you can try Julia.

Personally, if I were ever to ‘actually deploy’ these things… I wouldn’t dare use Python, I’d do C/C++.

There is just too much overhead in typing and it is way to slow. Very many of the common ‘libraries’ you use in Python are basically written in either of those two languages anyways.

For personal projects, that is fine… But ‘deployment’ ?

I always enjoyed this clip from the late, great (discoverer of the ‘bug’) Rear Admiral Grace Hopper:

I don’t know about you, but in production… I’d need my nanoseconds :wink:

2 Likes

Thanks for all the information, let me google what Julia is. :smiley:

Grace Hopper… what a beautiful mind!