Re:R in Ecology
by Wonnk13 on Wednesday February 24, 2016 @09:40AM
(#11167016)
Wow, i know nothing about Ecology, but used to be pretty proficient in R. I don't recognize half those libraries. I think R can pretty well be divided into pre and post "Hadley Wickham" eras. It's really a new language.
Re:R in Ecology
by hadley on Wednesday February 24, 2016 @01:41PM
(#11169060)
It's a bit more complicated than that - dplyr and magrittr invented pipes independently at about the same time. I preferred magrittr's implementation, so switched to that. dplyr continues to import and re-export %>% so that you don't need to explicitly load magrittr.
Re:R in Ecology
by muhic on Wednesday February 24, 2016 @02:54PM
(#11169690)
I wasn't aware of this parallel origin for the pipe. Just out of curiosity what did magrittr's implementation have over dplyr's? Thanks for incorporating forward piping in dplyr, it improves functional order/readability by miles!
Re:R in Ecology
by larrydag on Wednesday February 24, 2016 @09:41AM
(#11167027)
As a statistician I believe this article shows the strengths of R
- fast prototyping methods
- easily repeat research with scripts
- data visualization
- huge repository of add-on libraries from many areas of study
Sure R has it setbacks. I wouldn't rely on it for a production environment or a substitute for a higher level programming language. Yet if you want to study data and get results rather quick there is few that could compete in this space.
Re:R in Ecology
by brudgers on Wednesday February 24, 2016 @11:19AM
(#11167856)
This is a guest post from Auriel Fournier, a PhD Candidate with the Arkansas Cooperative Fish and Wildlife Research Unit at the University of Arkansas.
A real life example of what Scott Hanselman calls "a dark matter programmer"...he uses "the chief software architect of the Nebraska division of forestry."
Re:R in Ecology
by brudgers on Wednesday February 24, 2016 @12:32PM
(#11168493)
Except for the past few weeks, I've listened to all of Hanselman's Hansel Minutes podcasts, started in summer 2014 and got current about a year ago. Interesting content while driving in my opinion.
Anyway, he uses the term fairly regularly as a way of grounding discussions when they drift toward the-newest-shinyest or what-the-cool-kids-use. I think it's a helpful construct when thinking about "what should I use" since the best answer usually isn't React Native.
Re:R in Ecology
by Demagog on Wednesday February 24, 2016 @11:58AM
(#11168194)
It is very weird to me, why he doesn't use more of tidyr and dplyr. Those are wonderful packages. The whole part of changing types of data frame, would become single call to `mutate` from dplyr package.
There is great blog [0] post explaining working with those tools. What is most important here, is that dplyr and tidyr makes data organization and analysis concise, effecient and readable way.
Re:R in Ecology
by pvaldes on Thursday February 25, 2016 @05:31AM
(#11173545)
I miss some important packages here (IMAO), but is an interesting entry to further readings.
Otherwise, I don't really understand the hype with that %>% thing and did not find nothing about this structure in a quick google search. I would really appreciate if someone can enlight this fool. Thanks.
At any rate, for those looking to do distance sampling/other ecological work, check out this book: https://github.com/dill/RDistanceBook and this site: http://distancesampling.org/
Actually, it was magrittr[0]. dplyr popularised their use.
[0] https://cran.r-project.org/web/packages/magrittr/index.html
- fast prototyping methods
- easily repeat research with scripts
- data visualization
- huge repository of add-on libraries from many areas of study
Sure R has it setbacks. I wouldn't rely on it for a production environment or a substitute for a higher level programming language. Yet if you want to study data and get results rather quick there is few that could compete in this space.
A real life example of what Scott Hanselman calls "a dark matter programmer"...he uses "the chief software architect of the Nebraska division of forestry."
Anyway, he uses the term fairly regularly as a way of grounding discussions when they drift toward the-newest-shinyest or what-the-cool-kids-use. I think it's a helpful construct when thinking about "what should I use" since the best answer usually isn't React Native.
There is great blog [0] post explaining working with those tools. What is most important here, is that dplyr and tidyr makes data organization and analysis concise, effecient and readable way.
[0] http://zevross.com/blog/2015/01/13/a-new-data-processing-wor...
Otherwise, I don't really understand the hype with that %>% thing and did not find nothing about this structure in a quick google search. I would really appreciate if someone can enlight this fool. Thanks.
http://blog.revolutionanalytics.com/2016/01/microsoft-r-open...
Edit: Just saying why the GP claimed Microsoft "owns" R, which they clearly don't.