data science for babies- a support group

[sandwitch]

I just enrolled in Coursera's Data Science Toolbox.

In this course you will get an introduction to the main tools and ideas in the data scientists toolbox. The course gives an overview of the data, questions, and tools that data analysts and data scientists work with. There are two components to this course. The first is a conceptual introduction to the ideas behind turning data into actionable knowledge. The second is a practical introduction to the tools that will be used in the program like version control, markdown, git, Github, R, and Rstudio.

It would be really great if anyone else wants to do this with me. I've signed up to many of these and never follow through, but I want this time to be different. Who wants to jump on the bandwagon?

[ferrus]

Well it's short and I've wanted to pick up some R for a while now. So I'll bite.

[ACow]

I don't believe I shall do the course, but I wouldn't mind discussions or I might even be able to help with any questions regarding it :P (I signed up anyway...there was an issue with the chinese version, but i managed to find my way over to the english one.)

Warning: I kinda do this for a living, and i'm a cantankerous/mystic sort, so I disagree with some of the material, but I swear i'll try to be as straightforward/non-cryptic/non-confusing as possible if I can help in any way, though I see ferrus is there, so I'm sure you'll be fine.

[ferrus]

Warning: I kinda do this for a living, and i'm a cantankerous/mystic sort, so I disagree with some of the material, but I swear i'll try to be as straightforward/non-cryptic/non-confusing as possible if I can help in any way, though I see ferrus is there, so I'm sure you'll be fine.

Are you much of an R wizard, out of interest?

https://class.coursera.org/rprog-016...26type%3Dlogin

This one seems to be more pitched at my level.

[sandwitch]

Warning: I kinda do this for a living, and i'm a cantankerous/mystic sort, so I disagree with some of the material, but I swear i'll try to be as straightforward/non-cryptic/non-confusing as possible if I can help in any way, though I see ferrus is there, so I'm sure you'll be fine.

I'm particularly fond of cantankerous criticism of coursework, so please have at it.

[ACow]

Are you much of an R wizard, out of interest?

https://class.coursera.org/rprog-016...26type%3Dlogin

This one seems to be more pitched at my level.

I would never describe myself as one (as in i still have to look up syntax, functions and research why the fuck bizarre error message/bugs came up), but i've also met people who do say they know R at work, and its fair to say i had a deeper understanding of what was going on than them.

I describe myself as an R tolerator, because i generally do think its a god awful language: its inconsistent, its too inefficient to do real large work in, and its dynamic/non-standard evaluation quirks means its dangerous to work with others/put into production because it basically just opens the door for undocumented bugs. But i'm not a fool and can see its got academic momentum behind it, its free, its got a huge library, and its main use theoretically should be for experimental/interactive small scripts. There's a bunch of non-programming stats students who are all coming out of university irritatingly evangelistic about it and they just keep repeating mantras about it, so what are you gonna do. Worse is better...ish.

But to me its really just a weird mash of interpreted lisp with algol-esque syntax, so I think I've got a bit more of an insight than most (being one of the five people left who still speak in lisp and don't find its functional features weird or confronting)...

Edit: Oh, and since 70%-80% of real world data science is data manipulation and transformation (and only then can you apply your models/algorithms), i feel its worth pointing out that things like python/commercial stats packages are still far better at that than R IMHO...

[ferrus]

I see. Well I use Matlab for a lot of the fuzzy logic work I've been doing recently and it seems to have its similarities. I do remember using SPSS at university and despising it irrationality.

[ACow]

Don't let me put you off, different strokes for different folks and all. And I'm known to be far more concerned with efficiency and design than most industry, let alone academic use. Even I'm realising I have to know more R.

Just for others, if memory serves me correctly, I believe I've looked at the r programming course ferrus is talking about, and iirc, i considered it best for people who already had some programming experience. Just checking, but assignment 2 is probably going to be a nightmare if you're not already familiar with a lot of the concepts used, and I'm skeptical that even completing it will really sink those concepts in. Most R programmers are not that advanced...

But I'm also up for discussing that course as well....