Comments on: D-Day in the Public Mind

By: Aaron Gurwitz

Aaron Gurwitz — Tue, 08 Jun 2004 03:30:35 +0000

What D-day has going for it as an historical event to be commemorated is the fact that it took place on one day, June 6, 1944. If historical significance drove commemoration, we’d see a lot more articles about the Battle of Stalingrad than about D-day, and, if we did, people would have a much better idea about how the Nazis were defeated. More generally, I wonder whether single-day events capture more than their fair share of attention—Pearl Harbor, 9-11, the Ides of March—and what effect this might have on how we understand history.

By: bryan

bryan — Mon, 07 Jun 2004 19:43:27 +0000

A propos R, there is I believe currently a project underway to make it easy to interface J http://www.jsoftware.com (the language I brought up in the APL thread as being one of the successors to APL) with R.

By: eszter

eszter — Mon, 07 Jun 2004 17:18:09 +0000

James – I’m pretty sure the Lexis Nexis data base Kieran used here includes articles only from the print copy of the NYTimes, which wouldn’t necessarily be affected by the Web. The issue Matthew raises is a good one and I agree with Kieran that it would likely be quite a headache to figure out, unless you could find someone else who’s already figured it out and use their data for normalization. But of course, finding such a person wouldn’t be that easy either…

By: James Joyner

James Joyner — Mon, 07 Jun 2004 13:52:35 +0000

The spike does seem to occur with the widespread use of the Web. Presumably, the amount of news is no longer bounded by the number of column inches available.

By: Kieran Healy

Kieran Healy — Mon, 07 Jun 2004 12:13:23 +0000

Are you sure there aren’t just more stories in the New York Times on everything? Yeah, there is that—if I’d spent more than 20 mins on this I’d need to find some way of normalizing the measure, though this is harder than it sounds.

By: Kieran Healy

Kieran Healy — Mon, 07 Jun 2004 12:10:28 +0000

D2 - R is terrific. I use it for all the statistical analysis I do. The learning curve is steeper at the beginning because it’s a command-line/programming application rather than a menu driven thing like SPSS, but it’s much better, more powerful and flexible in the medium to long run. It’s well supported by a very active and smart community of users. Here’s a post href=”http://www.kieranhealy.org/blog/archives/000109.html” rel=”nofollow”>http://www.kieranhealy.org/blog/archives/000109.html of mine saying a bit more, with a bunch of links to the various books available about R. Here’s a screenshot href=”http://www.kieranhealy.org/blog/images/r-snapshot.html” rel=”nofollow”>http://www.kieranhealy.org/blog/images/r-snapshot.html of it running on my linux workstation, and here’s one href=”http://www.kieranhealy.org/blog/archives/FinderScreenSnapz001.html” rel=”nofollow”>http://www.kieranhealy.org/blog/archives/FinderScreenSnapz001.html of it on my PowerBook. R works best with Emacs and LaTeX because Emacs Speaks Statistics href=”http://www.analytics.washington.edu/statcomp/copy_of_news/projects/ess” rel=”nofollow”>http://www.analytics.washington.edu/statcomp/copy_of_news/projects/ess and R supports literate programming methods href=”http://www.ci.tuwien.ac.at/~leisch/Sweave/FAQ.html” rel=”nofollow”>http://www.ci.tuwien.ac.at/~leisch/Sweave/FAQ.html that make doing reproducible data analyses really easy. But its console application is fine and other editors support it as well. A finance guy like yourself will want to check out RMetrics href=”http://www.itp.phys.ethz.ch/econophysics/R/” rel=”nofollow”>http://www.itp.phys.ethz.ch/econophysics/R/, a package of functions for financial engineering and computational finance. I know nothing of this. There is also RQuantlib href=”http://dirk.eddelbuettel.com/code/rquantlib.html” rel=”nofollow”>http://dirk.eddelbuettel.com/code/rquantlib.html. R’s main advantage is that it’s developed by really top-notch people who have a best-practices approach to the functions and packages available for it. This allows schmucks like me to avoid mistakes and appear cleverer than we are. I’ve used WinBUGS before, though not recently. (BUGS is short for “Bugger Windows has crashed/gotten infected/reported me to the RIAA again.”)

By: Matthew

Matthew — Mon, 07 Jun 2004 10:13:23 +0000

Are you sure there aren’t just more stories in the New York Times on everything? In Britain the (to give an obvious comparision) Times has gone from about 12 pages in the 1950s to more than a 100 pages today. Obviously most of this is features, not news, and I know it’s not as true with the NYT, but this could explain the gently upwards slope, no?

By: dsquared

dsquared — Mon, 07 Jun 2004 10:03:13 +0000

Is R any good, Kieran? Someone told me you can get WinBUGS (the “Win” is short for “Winners don’t use Macs”) to work with it these days, so I became interested.