collected R code

analyze stuff: Statistics meets rhetoric: A text analysis of "I Have a Dream" in R

Wow. I like this: Thomas Rahlf is writing a book (to be released soon) on how to create graphs in simple R. According to this post on R-bloggers, all code will consist of less than 40 lines. Some of the graph on the accompanying website already give me an itch to try it out myself. I hope he's going to put up the code on the webpage as well (website says so, available from January on). And while quite possibly not buying the book myself, I'm going to recommend it for our lab.

No kidding! Also, Paper available in elvish, dwarfish, but not in babelfish. HadCM3L ...one model to bind them all!

fortune("friday") #Thanks, Dirk: it could. Have a nice weekend, everyone.

I might add I learned some things: df$coloumname and df[['coloumname']] give you the same result; apply does not work on data frames; data frames are actually lists; sapply works on list; every actor has a Bacon Number. Quite something for a Friday. Now, back to vectorising a for-loop.

(Orginial plotting method from vegan_2.0-9, courtesy of Jari Oksanen.)

Found the above link on stackexchange - yay, I'm not alone. However, I'm hoping to see a solution. Can't find one, going home now. Life's short, and night fell already. Help appreciated! [Edit: I know, I can just save it in the correct encoding via RStudio "save with encoding". It would be nice, though, if I could just call the re-encoding from the console; however, I get another "Status 1" error.]

Lately, The Daily Mail (sic!) brought us a tabloid (...cough...) of georeferenced ancestry data for the US. It hit British newspapers via the web, that is: Upworthy, which, in turn has it from the US Census Bureau. If you ask yourself why I'm reblogging this: you can get the data and do this map yourself - in R, of course. Or plot other stuff. Like, totally superfluous maps. Or, like, real nice office posters. I'm just pinning this here not to forget these links. If I have free times at my hand, I'm likely doing an overlay of population density with the ancestry data. Like, totally, really. Did I mention I rather like the US approach to this federal data? It's available, because of the law. That's much better than most other countries, where you have to beg for most of the data and it's still not available - well, because of, ehhhr, Reasons. With a capital R. (Not to confuse with the software.)

And in addition to Word: there's LaTex, too. This post of Tal Galili should help here.

It's a pain that we need MS Word at all, but it won't go away. Editors, reviewers, co-authors and everyone else is using it. So: we have to manage to get tables out of R. Special pain: align to the decimal point. Found this here, which has some useful advice - but I have not tested it yet, so beware.

DF <- DF[which(!is.letter(as.character(DF$factor))),]

Biology, and especially evolutionary biology, produces the perhaps most influential metaphors and images of our time. People often get it wrong, but these authors here are walking on a very sophisticated thin line: they make use of the Diversity-Stability hypothesis, of the Intermediate Disturbance Hypothesis, and of equilibrium state assumptions or maybe even Equilibrium Theory. I think they read their share of papers and are not referring to naive assumptions of the "balance of nature". To make my point: the hypotheses mentioned are still under discussion, yet they develop a strong impact in other disciplines. What we are doing as biologists does matter. However, we are not in control what people will make of it.

What we are doing is rather desperate: "Can I use your data?" Because it's Friday. Because it's 2013. Because it is still true. Because this is not going to change easily. And: because it's worth a laugh. Everybody should watch this. At least everyone who wants to stand on the shoulder of giants.Or wants to check if something was done properly. (BTW, are Reinhart & Rogoff already a meme in the stats community?) (via. I recommend visiting.)