rstats-wtf / wtf-2019-rsc Goto Github PK
View Code? Open in Web Editor NEWWhat They Forgot to Teach You About R, 2019 January 15/16 @ rstudio::conf
Home Page: https://rstd.io/wtf-2019-rsc
What They Forgot to Teach You About R, 2019 January 15/16 @ rstudio::conf
Home Page: https://rstd.io/wtf-2019-rsc
I think this could be really helpful. I would have liked to spend more time on the different options that we could set here rather than making animals talk on start (although this is cute). What are the options that Jim, Jenny and other RStudio people set in their .Rprofile?
The project-centric workflow covered is a huge is a pre-requisite to enable life in the ephemeral cloud, where compute resources are created and destroyed on demand.
It would be terrific to explore the gap between the content of this workshop and being productive in the cloud where naked machines need to be dressed with R and the R project before analysis can begin.
Jim's slide 26 @jimhester
fs::dir_create("~/Library/R/3.5/library")
Would be nice to have a few .Rprofile .Renviron example files added in the slides
give examples about interactive vs non-interactive R mode would help too.
What is the role of the personal access token (PAT) in relation to the two authentication methods available on GitHub (SSH/HTTPS)? Is it only used to replace the password in HTTPS authentication? If so, what are the relevant differences between using SSH vs HTTPS with a PAT?
In this lesson we dove right into the weeds of R, and I wasn't clear on why we were learning this or how it would be useful in the larger context of the optimal workflow that this course is about.
Put an emoji reaction on the relevant comment below to indicate which operating system you use.
You have a set of tests: A, B, C, D
Each test has subscales: X, Y, Z
At each level of test, you want to score all subscales, so
A: X, Y, Z
B: X, Y, Z
. . . and so on
Jenny suggested that instead of thinking of this as an iteration problem, it's best to wrangle the data so that the two columns test
, subscale
are unnested into a single column
Leave emoji under a comment
The first bug was a type issue. Defensive programming would have made resolving this bug a non-issue.
@jimhester mentioned several options to consider: assertive, assertthat, checkmate, stopifnot, etc.
where: https://jennybc.github.io/wtf-2019-rsc/ under Schedule, Wednesday
current text: 2_2: Personal R Administration 1 of 2
suggested change: 2_2: Personal R Administration 2 of 2
Here's what I'm talking about:
https://rstd.io/wtf-2019-rsc
We're using GitHub Pages, in it simplest form, to create a landing page for the workshop, coupled with an rstd.io short link.
We want to be lightweight and nimble, because we always need to make live edits. We want minimal moving parts.
Not sure what the "best" theme is. But I know I'd rather get happy with a canned theme than do any Jekyll config.
I'll keep playing with this and am happy to receive suggestions.
The default theme looked bad, because the landing page had two headers. Possible solution:
https://github.community/t5/GitHub-Pages/Remove-Heading/td-p/4220
I have mixed feelings about getting into Jekyll config for this.
Update: I've switched to a different theme, where this bugs me less.
Do you think you all can touch on how/when to ease the transition from, for me at least, what usually starts out as a pretty free-flowing and chaotic EDA on a new project, to the type of scripted and controlled workflow that was covered yesterday? My research projects tend to be pretty unstructured at first, and even though I use version control, can get a bit bloated as I try new analyses. If when we settle on a workflow like that covered yesterday, what do we do with all the other code? Keep it in a repository, and start a new, smaller and refactored one?
You mentioned it a bit with the smell-test.R file, but I often end up with an entire smell test repo. Should these all be on branches?
Cheers
At work, I'm working multiple reports that all rely on common data files that occasionally update. Originally, each report lived in its own project with its own Data folder. I then ran into the issue of not being able to easily track which projects had the most recent data. As a result, I migrated them all to one my_reports mega-project with a single Data folder. Now I have trouble organizing my files in a way that follows project-oriented a workflow. Is there a way to have both reliable data and a good file structure?
Create the Gitter room and replace the link placeholder in the README
After running usethis::use_course("rstd.io/wtf-debugging")
, we get a directory of R scripts and a CSV file.
Suggestion: these files might want to be stored in a project so that it is self-contained.
Is there a way to do this without rendering a .rmd
? I found this code but couldn't figure out how to apply it:
tempDir <- tempfile()
dir.create(tempDir)
htmlFile <- file.path(tempDir, "index.html")
# (code to write some content to the file)
viewer <- getOption("viewer")
viewer(htmlFile)
@jimhester or a TA:
In session 3, we will fork and clone this repo.
Then I want a change (see below).
Then we will add this as upstream remote and pull this change in.
Please remove this line (or wherever it has ended up):
https://github.com/jennybc/wtf-2019-rsc/blame/master/day2_3.md#L25
and include "closes #15" in the message (also good to demonstrate that!)
The pre-work recommended getting the latest versions of R (3.5+) and RStudio (with no minimum version mentioned). It seems that having the Terminal (available in 1.1+) is critical; can it be made clear that is the oldest viable version for this workshop?
I get that ideally we'd all be working with the latest versions, but for those of us stuck behind layers of IT bureaucracy it's not always so simple.
What sorts of things should be in .gitignore beyond the defaults? If it's something like a whole directory (say, Data), can that be done with something like regex or do the files need to be added individually?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.