That no one knows it. For many ML algorithms they rather get published first in R than in Python still, although I believe this is not true for deep learning, but rather stuff like tree based models.
Sometimes I can’t implement what I want to because I don’t have time, let alone skill to port a library from R into Python and make it sklearn compatible. But I also don’t want to be the only person in my team who understands the code. So Python it is.
Edit: just realized this is 7 months old, but I guess maybe we can boost activity by also answering to old posts?
The move to Linkedin seems a bit weird but I’m glad that at least they’re also branching out to Mastodon, both are def better alternatives than fucking twitter
Hello, I am an IT support tech/sysadmin who somehow ended up with data analyst/engineer duties.
I develop R scripts (tidyverse), but I’m also responsible for running them as part of my team’s workflows. The scripts are basically manual data pipelines since the higher ups won’t let my team invest in actual automation & integration. I also use Javascript/HTML/CSS to build front ends for our information management system.
I don’t know many R programmers IRL so I’m glad to find a forum for it in the fediverse.
I am a an engineering consultant, specializing in solar photovoltaic power systems.
Tidyverse and Base R. I find people who do things the hard way tedious just to be pure.
I suppose I am more of a user, because I haven’t released any of my packages.
Almost 25 years.
C++, Python. Many other languages as they have been needed, from Assembly to VBA and Matlab.
R is to me what Excel is for a lot of people… a full featured calculator. I compose a lot using RMarkdown/Quarto. I tend to build reproducible pipelines for data or simulations.
Both, though at work there is some pressure to use more Python.
I have some stickers. I don’t have many.
Probably closer to the data wrangling end.
onion… 3d space rotations made easy through obscure math (which way is the solar panel pointing anyway?)
Re: c) I will be a dirty shill for VSCode and R lol, example here. I find it much better for R shiny development, projects with multiple people and projects with multiple languages. Notebook support is less good out of the box, you will have to get a jupyter kernel set up - but I use scripts more so than notebooks anyway.
Anyway, onto the question! Base R. Yeah, I said it! Whenever I have a weird enough situation where tidyverse functions won’t work due to poor quality data, then I shed a single solemn tear and quietly wish I had done the project in python as I start writing a for loop in what will no doubt be the most hacky solution ever.
As far as I know, yeah, R always work with copy on modification. Some libraries as you mention (data.table) can have object/classes to avoid this, but I’m not aware of any of them working with arrays (more than 2D). Maybe parquet or arrow have something like this??
+1 for parquet and arrow. If you’re pushing memory better to just treat it as a completely out of memory problem. If you can split the data into multiple parquet files with hive style or directory partitioning it will be more efficient. You don’t want parquet files too small though (I’ve heard people saying 1 GB each file is ideal, colleagues at work like 512 MB per file - but that’s on an AWS setup).
Bonus is once you’ve learned the packages it’ll be the same for all out of memory big datasets.
I’m @MalditoBarbudo, a data scientist at an ecology and forestry research center.
I’ve been using R for 14 years, starting as an user and ending as a developer. I’ve done also some python, sql and web development (html, js and css).
I prefer tidyverse, it fit perfectly with my mental logic, but I reckon that sometimes data.table is needed (but in that case dtplyr comes to help!).
I maintain several packages (sapfluxnetr, meteospain…) and collaborate in others (meteoland, medfate…). I also maintain a web with several shiny apps for forest data visualization (LFC).
My favourite obscure package changes every week or so, but if I have to choose one, lately I’ve been playing with rayshader, trying to create nice 3d map plots.
I've been reading books about GLM and Bayesian Statistics and am using the "glm" function and "brms" and "rstanarm" packages. They're pretty fun. If any of you are interested in using those packages know that rstan is a bit weird which is why I'm using R 4.0.2 and rtools42 and not the current versions.
r_programming
Hot
This magazine is from a federated server and may be incomplete. Browse more on the original instance.