Question

I'd like to learn data analysis/stats and perhaps modeling... I'm looking for an online course, or your favorite free information source.

My background is vet medicine and I'm interested in fish biology, so if possible, I'd like a bio-focused source. I'm looking for a masters or phd-level resource.

Thanks!

Answer 1

Okay cool! That happens to be a unit I'm panicing about right now, which ironically means I happen to have a long list of things you could use. I would start here:

http://cran.r-project.org/other-docs.html

R is a statistics program used widely in the biological sciences - so most of the tutorials on there will help you with both :D

Answer 2

Trance is absolutely right: Real Biologists Use R.

Virtual high five given.

There are a plethora of good online tutorials specifically for R, both written ones and videos on YouTube, which cover not only using R but the statistics underneath. I am slowly and quietly putting together another, better bio challenge, one which will need R - attached is the intro to R I wrote for that. Please go through it and any feedback you would like to give me about it, I would definitely appreciate.

Answer 3

@blues and @TranceNova Thank you! I wasn't even sure where to start -- taking the coursera modeling class through U of M and it's simplistic and aimed at undergrads.

My thanks!

Answer 4

Great! Hope you enjoy it and if you have any questions don't hestitate to ask :D

Answer 5

Hope you find the intro script helpful.

How is model thinking, btw?

I'm in Coursera's probabilistic graphical models from Stanford. 10 week course, 1200 page book, half the Ph.D. students who try it either drop or fail - no more free time.

Answer 6

@blues Model thinking class is so far a bit fluffy. It's not awful and I don't quite feel like it's a waste of time, but I wish I could fast forward about 50% of the material - however, I'm only on the "second lecture" (the videos are ~10m each so many comprise a lecture), so that may improve deeper into the course.

Yow! That one doesn't sound fun at all! How is it, besides overwhelming?

Answer 7

PGMs is pretty sweet. But I like to stay on my toes.

She moves quickly. Generally 15~20 min lectures, 2.5~3 hrs per week. The problem sets are OK but the programming assignments. Those have been time consuming. But it is sooo interesting!

Answer 8

After running into a problem figuring out how to *run* R ('type 'R' in terminal window' is the answer), I've managed to run through the first quarter of @blues' file, above, and it's working well!

Reading ahead, there seems to be a jump in assumed comfort of the reader just after ds is introduced. I can provide more details after working through it tomorrow, if you'd like?

A more general question - how did you learn beginning R? Where did you get questions/sample data? I learned C++ a million years ago in school and this seems to be a similarly open-ended hugely powerful but amorphous language. I had trouble then figuring out what I wanted to make a program to do and I suspect I'll have the same problem w/ R... I assume it's mostly learned within a class context or when you have data in need of analysis and you're the only one to do it?

Thank you both!

Answer 9

Well I started learning R this year (so not too long!). I have a unit that's for ecological statistical analysis, so at the moment that's where I'm getting my data and questions from. I think R is more orientated towards biologists and other non-programming people but you can code are in more "programming" way - which is good if that is your background. R has heaps and heaps of functions which is pretty widely applicable, it also has a large userbase so most things are pretty "google-able" if you need to work how to do something.

I learned it in a class context but I'm a postgrad student and the course is aimed at those of us that have our own projects. So later this year I will be using to analyse my own ecological data.

Answer 10

So, it sounds like R is what you need... However, I don't model the types of things it does. I'm doing cancer research using Computational Molecular Biophysics. In that field, 90% of the modeling comes from Autodock, and inputs are made using Autodock Tools.

http://autodock.scripps.edu/

As for the statistics, we do ours by hand... aside from the Root Mean Squared Deviation that autodock generates.

Answer 11

"As for the statistics, we do ours by hand" I think you scared me a little there :P

Answer 12

Well, we mostly do it by hand. There have been a few times we used Mathematica to generate results (because of the number of entries) and to create tables, etc for publication.

http://www.wolfram.com/mathematica/