Monthly Archives: June 2010

ggplot2 GUI progress

June 19, 2010
By

(Written by Ian Fellows) Below is a link to the first of a weekly (or bi-weekly) screen-cast vlog of my progress building a GUI for the ggplot2 package. http://neolab.stat.ucla.edu/cranstats/gsoc_vlog1.mov comments and suggestions are more than welcome, and can e-mailed to me at: [email protected]

The perfect fake

June 19, 2010
By

Usually when you are doing Monte Carlo testing, you want fake data that’s good, but not too good. You may want a sample taken from the Uniform distribution, but you don’t want your values to be uniformly distributed. In other words, if you were to order your sample values from lowest to highest, you don’t

Why R doesn’t suck

June 19, 2010
By

I first encountered the R programming language a few years ago when I needed to make some plots. Although I’ve used it occasionally since, I always considered it a sort of “Perl for statisticians” — a useful swiss-army knife with … Continue reading →

Those dice aren’t loaded, they’re just strange

June 18, 2010
By

I must confess to feeling an almost obsessive fascination with intransitive games, dice, and other artifacts. The most famous intransitive game is rock, scissors, paper. Rock beats scissors.  Scissors beats paper. Paper beats rock. Everyone older than 7 seems to know this, but very few people are aware that dice can exhibit this same behavior,

Revolution Analytics: Startup to watch

June 18, 2010
By

Jack Germain of LinuxInsider interviewed Revolution CEO Norman Nie for his "Startup to Watch" column. Amongst the topics covered: the R language (Norman: "There are no statistical expressions that can not be written in R"), Revolution's recent name-change and announcement of our development roadmap, and the challenges of competing with SAS and Norman's former company, SPSS. Read the full...

The impact of the drug war in Mexico

June 18, 2010
By

For the last couple of years, Mexico has been in the midst of an escalating drug war, with violent crime on the upswing in many areas. But tracking the impact quantitatively is difficult: in Mexico, about 85% of crimes go unreported, and corruption leads to inaccurate reporting in some districts. Diego Valle has taken on the task of visualizing...

June 18, 2010
By

R Commander – linear regression

June 18, 2010
By

We can fit various linear regression models using the R Commander GUI which also provides various ways to consider the model diagnostics to determine whether we need to consider a different model. Fast Tube by Casper The “Statistics” menu provides access to various statistical models via the “Fit models” sub-menu including: Linear regression – the simplest scenario with