Tyler Littlefield

scale_fish_viridis(option = "plasma") #rstats #viridis

@MaxRovensky @dhh @heyhey Same, I can’t use this app anymore. I’m literally in my home with 2GB WiFi and still can’t read an email.

@dhh Ever since you started "moving out of the cloud", my experience with @heyhey has increasingly been this

"After six months of effort, it's done. We're out of the cloud. The last application was brought home to our own hardware on Wednesday. Hallelujah!" world.hey.com/dhh/we-have-le…

@MikeCalcagno @heyhey @37signals Happens all the time for me. Worst part is every time it happens, I try another mail client like native Apple mail app and it works.

For all the trash talking that @heyhey and @37signals do, this happens way too often. Like, daily.

@koen_hufkens @jakub_nowosad @TrevorABranch While renv is a great initiative, I have honestly given up on it due to issues with cross platform / system dependencies. I just use docker containers these days. Maybe I am doing something wrong, but everyone on my team hates me for introducing renv.

So much for all your R code always being reproducible. Seriously, this is a major issue. Most scientists use R for a project and expect their code to work when someone tried to reproduce what they did. Removing older packages from CRAN is not a great solution.

#rstats when using multiple repositories (external/internal), can I specify repo priority? For example, always try to install from internal repo, then fall back on external.

@Rami_Krispin Do you have a strong preference on volumes vs bind mounts? Curious as I am having my team switch over to local rstudio-server via docker and persistent storage is the main thing I want to get right

Blog post I am referring to at: loppsided.blog/posts/2020-10-…

Initial attempt at CML + tidymodels to display model outputs/metrics with every commit! Inspired by a blog post on CML/DVC by @lopp_sean #rstats github.com/tyluRp/mlops/p…

I have an lapply that sources some R scripts, which in turn seems to make memory usage creep higher and higher. If I use a for loop instead, does it magically prevent the memory from growing? Can anyone enlighten me on whats going on? #rstats

@sean_turner @rstatstweet Thanks, I'll give it a shot

@tyluRp @rstatstweet I had all sorts of debugging issues in r studio of late. Resolved by updating to the development version.

@dgkeyes Consistency in business logic. We often have to do a repetitive workflow across multiple datasets maintained by multiple people. An internal package allows us to package, maintain, and test logic that is reused consistently by everyone on the team.

For those of you who work at organizations with internal R packages, what are the main things those packages do? #rstats

Finally giving #PopOS a try

@mdneuzerling @erdirstats Doesn’t add authentication to the api, but wanted to share this cool package for rmd authentication: mattwarkentin.github.io/fidelius/

If not, I am really just looking for a safe way to replace a table with new data. And by safe, I just mean that while I replace the table, the "old" data is still accessible to the user until the "new" data is available.

Can I use pg_dump within {dbplyr}? #rstats

@matherion @rstatstweet Yep! It's hundreds of powerpoint slides basically, split by specific categories so it was a must have. None of us wanted to spend hours making it by hand 😅

Am I doing this correctly? I want a higher level function that calls on lower level functions and allows users to provide any parameter for those lower level functions #rstats

@matherion @rstatstweet Ended up going with this

@matherion @rstatstweet Hmm, I guess I would be nervous if child functions had the same parameter name but expected different values. Whether or not that is good API design is another topic hahaha but just something I thought of.

@matherion @rstatstweet Happy you like it!