Twitter | Search | |
Joe Thorley
A computational biologist who programs in R and fits models using JAGS (BUGS) and TMB (ADMB).
780
Tweets
94
Following
122
Followers
Tweets
Joe Thorley retweeted
InteriorFraserSteelhead Nov 6
Reply Retweet Like
Joe Thorley retweeted
Hadley Wickham Nov 5
My take on why you should use purrr::map() instead of base::lapply() — it's not about one function —
Reply Retweet Like
Joe Thorley retweeted
Richard McElreath Nov 2
Replying to @rlmcelreath
Deep relation to maximum entropy interpretation of probability. Steve Frank has a great paper about this from 2009:
Reply Retweet Like
Joe Thorley retweeted
Richard McElreath Nov 1
New Statistical Rethinking (a Bayesian Course etc) lectures for Fall 2017. Lecture 01: The Golem of Prague (60mins)
Reply Retweet Like
Joe Thorley retweeted
seb dalgarno Oct 31
Paper from my M.Sc. thesis published today in ACE on Black Oystercatcher distribution modeling in Haida Gwaii.
Reply Retweet Like
Joe Thorley retweeted
Lauren Brent Oct 31
Very appealing visual Introduction to hierarchical (i.e. mixed effects) modeling by .
Reply Retweet Like
Joe Thorley retweeted
Garrett Grolemund Oct 27
New cheatsheet today: Strings with ! Here: Thanks for diagrams and inspiration.
Reply Retweet Like
Joe Thorley retweeted
Poisson Consulting Oct 30
Shiny webpage to view the tides at hundreds of US stations
Reply Retweet Like
Joe Thorley Oct 29
Replying to @big_bad_sam
Also worth noting that distinct at the start facilitates autocompletion
Reply Retweet Like
Joe Thorley retweeted
Andrew Gelman Oct 29
“Quality control” (rather than “hypothesis testing” or “inference” or “discovery”) as a better metaphor for the stat
Reply Retweet Like
Joe Thorley Oct 28
Replying to @joe_thorley
better yet library(dplyr) library(rlang) select(mtcars, UQ(parse_quosure(“disp”))) mutate(mtcars, new = UQ(parse_quosure(“disp^2”)))
Reply Retweet Like
Joe Thorley Oct 28
FWIW dplyr::select_(new_name = ~term) can be replaced by dplyr::select(new_name = !!(term))
Reply Retweet Like
Joe Thorley Oct 25
Everything is connected to everything else - Commoner 1971 thus with sufficient n any relationship is significant!
Reply Retweet Like
Joe Thorley retweeted
Luke Thompson Oct 24
All the more reason to focus on the systematic forces behind environmental destruction rather than individual behavior.
Reply Retweet Like
Joe Thorley Oct 20
Replying to @big_bad_sam
I go with as short as possible without compromising understanding (and definitely not more than 80 characters ;)
Reply Retweet Like
Joe Thorley Oct 20
This is fascinating both statistically and ecologically
Reply Retweet Like
Joe Thorley Oct 19
Replying to @joe_thorley
Fixed as of v1.10.4-3!
Reply Retweet Like
Joe Thorley retweeted
Alain Zuur Oct 19
Course in Banff, Canada: Intro to regression models with spatial & temporal correlation using R-INLA. See
Reply Retweet Like
Joe Thorley Oct 18
Parallelism, R, and OpenMP - an amusing (at least to me) introduction
Reply Retweet Like
Joe Thorley retweeted
Andrew Gelman Oct 18
No tradeoff between regularization and discovery
Reply Retweet Like