Data Science. TileDB. Open Source. Quant Research. R. C++. Debian. Linux. Adjunct Clinical Professor, University of Illinois. Lots of coffee. And some running.

Chicago, IL, USA
Joined March 2007
Filter
Exclude
Time range
-
Near
New blog post: Faster creation of binomial matrices, from an R speedup exercise started at the REcology blog goo.gl/TglMG #rstats
1
3
Replying to @jaredlander
@jaredlander Sorry, data not aggregated that way, no time to do it. In past, I created similar growth charts using public info off CRAN.
1
Replying to @jaredlander
@jaredlander The underlying db is not exported.
1
Replying to @petersagal
3:13 here MT @petersagal Yes, I took this opportunity to brag about my marathon PR. We all do that. But we don’t lie! This is not fishing!
Replying to @kevin_ushey
MT @kevinushey @rstudioapp RStudio now with build tools easing Rcpp development, C++ syntax + error parsing, Roxygen and more. #rstats
1
Replying to @JustinWolfers
@JustinWolfers 4:01 is a 9:11 min/mile, 3:00 is 6:52 min/mile -- a huge margin that is. We all celebrate shaving 10 sec/mile off.
1
Replying to @brendan642
Uses Rcpp so must be good :) MT @brendan642 Stan 1.0 MCMC framework is released! Looks pretty impressive from the 188 page manual. #rstats
Wow: Real Madrid, Manchester City, Ajax Amsterdam and Dortmund -- what a Champions League "group of death" that is. #soccerIsLife
Two very good points (on models vs truths, and open code) raised by @robjhyndman : goo.gl/5pPXj #rstats
2
Replying to @mikedewar
@mikedewar Not really. Probably neither first nor last time that non-information got dressed up in shiny bar charts. At least no pies...
Replying to @mikedewar
@mikedewar So a score of 108 with the other groups just below 100 seams to imply ... that this is totally uninformative. 2/2
Replying to @mikedewar
@mikedewar I think you are reading that wrong: "Index == 100" user as likely as random internet user. Index == 200 twice as likely. 1/2
1
Chapeau: Code from the parser package by @romain_francois just got included into R-devel goo.gl/BPWN4 #rstats
1
1
Replying to @RLangTip
And more complex code too @RLangTip You can use the Rcpp package to convert simple R loops into fast C++ code: bit.ly/RJLJpS #rstats
1
1
5
Replying to @mikedewar
weatherspark is pretty nice @mikedewar "What's the most beautiful weather report you've seen online?"
1
@olmjo: Rcpp calls through to R, so should be similar; other libs can win as eg GSL's rgamma is faster, cf old blog post #rcpp #rstats
Replying to @geospacedman
@geospacedman That was my phone autonomously tweeting while I was commutting.
Replying to @hadleywickham
@hadleywickham It is there to remind the young ones that there used to be another implementation of the S language :)
@mt_morgan @jmw86069 That's why I posted source code :)
1