Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

R Weekly Highlights

R Weekly Highlights artwork

Issue 2026-W19 Highlights

May 8, 202644 minAudio

If Shiny wasn't awesome enough, a new toolkit is bringing the awesome to new heights! Plus a handful of ggplot2 functions that can immediate impact common visualization operations, and knitr gains support for a much-hyped typesetting format for the static document crowd. Episode Links This week's curator: Eric Nantz: @rpodcast@podcastindex.social (Mastodon) & @rpodcast.bsky.social (BlueSky) & @theRcast (X/Twitter) Announcing shiny.webawesome: a web UI package for R/Shiny Five ggplot2 functions I wish I’d known about earlier Preliminary Support for Typst in knitr Issue available at rweekly.org/2026-W19 Supplement Resources Add links discussed in the episode (in place of this sentence) {shiny.webawesome} documentation https://www.shiny-webawesome.org Web Awesome: The Web's Open Source Design System https://webawesome.com {lemon} Freshen up your ggplots https://github.com/stefanedwards/lemon ligt.js Miscellaneous lightweight tools and utilities in JavaScript https://github.com/yihui/lite.js/blob/main/js/pages.js R 4.6.0 binary transition https://github.com/eddelbuettel/R-4.6.0-binary-transition Exploring and Learning Quantum Computing Part 1 https://www.kenkoonwong.com/blog/qc1/ Mike's talk "Fully Reproducible Model Validation Reports Using Docker, R, Quarto, {renv} and {targets}" https://youtube.com/watch?v=hRIbxv516hY Supporting the show Use the contact page at https://serve.podhome.fm/custompage/r-weekly-highlights/contact to send us your feedback R-Weekly Highlights on the Podcastindex.org - You can send a boost into the show directly in the Podcast Index. First, top-up with Alby, and then head over to the R-Weekly Highlights podcast entry on the index. A new way to think about value: https://value4value.info Get in touch with us on social media Eric Nantz: @rpodcast@podcastindex.social (Mastodon), @rpodcast.bsky.social (BlueSky) and @theRcast (X/Twitter) Mike Thomas: @mike_thomas@fosstodon.org (Mastodon), @mike-thomas.bsky.social (BlueSky), and @mike_ketchbrook (X/Twitter) Music credits powered by OCRemix Moonlight Vibin - Megaman X5 - DCT - https://ocremix.org/remix/OCR02053 Countdown to Infinity - Megaman X: Maverick Rising - WillRocky - https://maverick.ocremix.org/music.php

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

Hello friends, we are back with episode two twenty five of the Weekly Highlights Podcast. This is the weekly ish show where we talk about the great highlights and sometimes additional resources that are available on this week's issue of R Weekly, which is right at rweekly.org. My name is Eric Nance and I'm delighted you're joining us from wherever you are around the world, where off the bat I will let you all know I am recording this episode on a brand new server and if there are any gotchas on there, we'll find out. Mike and I had some fun banter about one little gotcha that's not gonna affect you all listening to this, but we'll we'll plow through. But, anyway, speaking of the of the man himself, Mike Thomas, as always, is joining me today. How How are you doing today, my friend? Doing well, Eric. I feel like we're getting a little bit more consistent, which feels good.

It does. It's been long overdue, but, yeah, I can pretty much assure you listeners out there that our release schedule, yeah, it's going to change a little bit from what it was before because it seems like early week for me has just been destroyed by my work calendar. So later in the week, we've been able to carve out some time on a regular basis and we'll do our best to adhere to this. So yeah, you'll get our weekly as usual, it's just a little bit later in the week, maybe a good way to start your Fridays

by the time this comes out there. But yeah, great to be consistent here and speaking of consistency, we got another batch of excellent resources to tell you about. And let's check my fancy note generator here, for who curated it. Oh yes, it is this old old guy here, Eric here, was a curator this week. Again, great resources to share. Merge many poll requests, which is always one of the best ways you all can contribute to RWeekly. Send us a poll request, we'll tell you more about that at the end of the show like we usually do, but as always, this is not just a solo effort for me. I had tremendous help from our fellow RWKY team

members, our curators, and of course contributors like all of you around the world of your poll requests and other wonderful suggestions. And in fact, I believe our first highlight today is indeed one of the poll requests that came through, and as Mike and I are very avid developers and fans of the Shiny ecosystem, this one definitely caught my eye and I'm eager to tell you all about it. What we're talking about here is the fact that in the Shiny ecosystem, there has never been a shortage of different, you might say, toolkits you can use

to bring like a different type of interface or user interface, I should say, on top of your application. We've talked about extensively since the advent of BSlib from the Shiny Team at POSIT, that's become a really big staple for the way I develop my applications in the current age and gives me a lot off the box or off the shelf to make my toolkit very responsive across different dimensions and a lot of nice enhancements that I know had a lot of attention to get there.

That's never been the only game in town because before BSLib, we of course had lots of great alternative dashboard layouts, especially those from a good friend of ours, David Grandin, with his amazing toolkits. But we have a new contender in town, folks, and it is called Shiny Web Awesome. And yeah, I think it lives up to the billing. This is pretty awesome from what I'm seeing. And Shiny Web Awesome comes to us from MB Anand, hopefully I'm saying that right,

and he or she has put this great blog post together that announces the package and gives us a little tour of what's under the tin here. So this is actually surfacing an upstream project called WebAwesome, which again we'll have a link to this in the show notes as well. It is a UI toolkit, you know, it's an alternative to Bootstrap and the like. Definitely has strikes the balance of modern widgets,

modern layouts, and some nice under the hood kind of extensions you can do with its kind of command interface. So right off the bat, which I like a lot here, this post gives us an example app right in the browser page via WebAssembly, of course, and you can see what this, you know, very utilitarian version of WebAwesome looks like in a Shiny app. And as I'm looking at this on my screen here, looks pretty familiar to most of the Shiny apps I see, where we got

kind of a nice header at the top, a nice rounded corner, so we're really making pop out, along with a sidebar with different input dropdown widgets, some nice radio toggles here, and like a button to pop up what's called an inspector right sidebar that comes right away, you can dismiss it, but it's a nice way to kind of hide additional options and maybe are more advanced. And then we have in the main panel on the right, a visualization of the IRIS data set,

looking with a, not sure if that's a ggplot2 plot or not, but we got three tabs here of the chart, the summaries, statistics where you can choose between the mean and median and whatnot, and additional details which look like kind of the expandable card or accordion panels that you can do in BSLib as well, but you can show and hide all those, as well to your leisure, and you get some interesting kind of call outs at the top of that with the number of rows, the correlation,

and what tab you're focused on at the moment, and some little buttons at the top that kind of reinforce that information as well. That looks pretty handy, and again, that can get started right away, the code is on the blog post as well, very few lines of code. If you're comfortable with shiny, yeah, I think you can feel right at home with calling these different container options they have to put like the different collection of widgets. You then do the typical UI and server kind of call

and you see all the parameters documented here as well. The rest of the post talks about kind of the design philosophy behind WebAwesome and this package itself, which is trying its best to stay as close to the upstream project as possible. I know sometimes that can be a challenge depending on which toolkit your package uses, but it looks like there are ways that he can keep in touch with, in essence, a scaffolding JSON or YAML file that the project maintains

so that he's able to simply source that and then link directly to the web assets from the awesome package, or the web awesome framework itself. There is support for custom shiny bindings for different events and reporting certain interactions in the app, which is great if you want to do some programmatic type of interactions, and he has links to the, package down site with these articles talking about things like shiny bindings and whatnot. So great, great, comprehensive examples that I see on the package site itself.

And then lastly, he mentions the command API, which might give you a little way to interact with additional server side functionality by using the WebAwesome kind of methods of doing either a call method or setting a property that lets you interact with a client side of the browser in more programmatic ways. You can even do custom JavaScript with a WAJS function

if you need even more customization to how you interact with the client side operations. I'm intrigued by this. I will definitely have a look at this for maybe some new apps that maybe are smaller scope in nature, but it does look like WebAwesome has a lot of interesting toolkits available here that you could choose for your next shiny app that you want to bootstrap. I think this is really interesting, and I hadn't even known about the WebAwesome project sort of outside of this Shiny context.

And as you mentioned, Eric, BSlib has had a huge footprint and impact on UI styling for Shiny apps in the past couple years, you know, particularly now with the advent of brand.yaml and our ability to sort of extend the same styling over Shiny apps, you know, Quarto documents, you name it. It was kind of a throwback to to hear you talk a little bit about, you know, some of the older UI styling, methodologies we used. It brought me all the way back to shiny dashboard.

Oh, yeah. That was an era. There were shiny dashboards absolutely everywhere. And I think it certainly paved the way for what we what we have now. It looks like from the reference page on the package down site when I'm looking at the the functions that are available, and it's a pretty extensive API here. There's kind of two ways to lay out your page. There's a WA underscore container function or a web awesome page function. So I think that's sort of your your base unit upon which to build. And there's lots of awesome components as well, including things like carousels

and breadcrumbs, which I think, you know, a lot of the time we don't get out of the box from, the shiny package or from BS lib. So that's pretty cool as well. I'm excited to explore all these different components. And the one thing that I really have to do a little bit more research on and I'm wondering is is how this relates to Bootstrap at all. I'm wondering if if WebAwesome is built on top of Boot strap or if those are kind of totally separate things or if Bootstrap is dead. Do you have any intuition on that, Eric?

I don't think it's going away anytime soon just given the prevalence online. But, boy, my first read of the WebAwesome library here itself, this looks like custom from the ground up. I don't see it sharing almost anything with Bootstrap here. Yeah. I agree from what I can tell. But I it might be worth putting in the, show notes here. You can go to webawesome.com Yep. To kind of see the underlying project itself, and it's described as the web's open source design system, biggest open source library of meticulously designed, highly customizable, and framework agnostic

UI components. So pretty cool to check out. I think it's fairly different than, in terms of like overall style. Maybe a little bit more playful than what you might see in bootstrap and the latest versions of boot strap. So maybe it tickles your fancy a little bit more than, you know, the traditional bootstrap themes do, which obviously underlie, what we see in in BSLib and Shiny currently. Right. And, of course, as you adopt these other frameworks, you are adopting kind of their own method of bootstrapping

these components and how you invoke them. At least from what I'm reading here, it should be if you have pretty basic requirements, I think it should be pretty easy to opt into. But, also, if you need some of these additional bells and whistles, then maybe BSWIP gets you kind of 90% of the way there, but you just wanna get to that extra bit. It might be worth kind of seeing what else WebAwesome gives you in this space. But again, choice is always good in the realm of developing your shiny apps, and I would be intrigued by seeing

what additional things we can do on the client side that maybe, A, are user driven or b) are more programmatic driven, like the breadcrumb feature. I think that's nice as you're navigating a more complex application and need to get, you know, somewhere back and you're not quite sure how to get there. So, yeah, just looking at the components on the WebAwesome upstream docs site, there are a ton of them. So yeah, I'm sure over time as this package matures more, you might get, you know, tailor made functions for these, and it does sound like you get a way to opt into the custom features

of your choosing if you find something that's not as supported right off the bat with an off the shelf R function in the Shiny WebAwesome package proper. So, yep, I will be filing this into my continued learning adventures next time I ask to bootstrap a new app, I will have this in my toolkit.

And up next, we're going to take a visit back to the visualization corner here, which is always fun to do on an audio podcast, but we learn a lot along the way as we do this. And when I, like, learn about visualization, there are, like, a handful of people in the community that when I see come through in an RWEQ issue, I'm like, yes, please. You get my time to read on this, and that is very much the case for this second highlight here because we have the esteemed Nicola Reeney who has returned to the highlights section once again. If we ever had a leaderboard about this, she would probably be near the top of the list, daresay,

but she has a new blog post on her revamped blog that we think looks fantastic called five g g pod two functions I wish I'd known about earlier. Oh, boy. Is that the story of my life of many of the functions I learn about, and I'm like, where was this? Or it was always there, and I just wasn't smart enough to look. Well, it's always good to highlight these lesser known gems, so to speak, in the world of g g plot two because we can do so much with that package. It's it it boggles my mind just the kind of creativity that's involved here.

So I'm gonna cover the first few here, then Mike will take over for the rest, but, we're gonna start one by one here. And the first function that she is using extensively now is the guide axis function. And where this can come to play is very much when you're entering the world of faceting in your plots. So in the example that she has here at the top here using the Gapminder dataset, It's a faceted plot based on country

where we have the year on the x axis and the life expectancy on the y axis. Very typical, you know, type of visualization you would see with Gapminder, but there is a problem with this default implementation after you're done with that facet wrap call at the end of the pipeline, is that the ticks or the year labels on the x axis, the individual years on the ticks I should say,

they are overlapping each other and good luck trying to read that even if you have a fancy four ks monitor like I do and I'm trying to zoom in on that, they're all overlapping. The best experience, especially if you're just exploring this data for the first time. Well, what you could do is you could go through the drudgery of adjusting this yourself, but guess what? There's an easier way, and that is simply after you're doing your scale call to tell,

in this case, that the x axis is a scale x continuous, you can simply feed in inside that scale x continuous a parameter called guide, and the value of that parameter is guide underscore axis, which you can put a host of different options here, but one of those that's really important is check overlap. And if you set that to true, well, guess what? When it outputs this plot and all the facets in this glory,

you will not see all the years output here. You're only going to see a handful of years that are actually labeled on the x axis. So yeah, this makes it a lot easier to read. Of course, you do lose as some of the ticks are just going to be blank, but I think if there's only about four or five of these levels, you can do the gist if you see the scale in between consecutive entries. So I think it's a worthwhile trade off and that way you get that readable plot with facets

on that first invocation once you do that guide axis call in the scale as continuous function in this case. Next, another area that can be a bit of a difficulty to deal with is when you're doing a plot and you want to annotate certain parts of that plot, such as points on a scatter plot, maybe a bar an indentation on top of a bar or whatnot. The text sizing always feels like a manual effort to get just right.

And so in the example here, we again have the Gapminder data set that's filtered to a particular year, she groups it by the continent, and then does a gets the max value of life expense expectancy and then does points on that with the life expectancy on the x axis and the continent on the y axis. And so you see the dots, and reason they got the max on there because that's where the annotations are going to go and the annotations are simply giving the country name,

but look at the text size in the post when you see this outside listening to us. She started with a size of 12 to make the text stand out, but it is a bit odd, And what you also notice is that the font size is actually not that really nice number, in fact I misspoke it's that, oh there's a size here, but the actual size itself is actually 3.88.

Well guess what, why is that? On the actual picture, the units of the font when you do the size column GM text is millimeters. That sometimes is a little odd, right? And what we usually expect is either units and points or maybe pixels depending on what annotation you're looking at. Well, guess what? There is a way to get a handle on that, and that would be by not just having the size supplied to GeomText,

it's the size dot unit, value, which you can choose then the different unit type you want to see for that particular size value, and in this example she uses the size unit of pt or point. When you do that, now you get what really looks like a font size of 12 here and not 12 millimeters as it was in the previous plot. So again, you won't see that by default, but once you know that once, I think your annotations

are going to look a heck of a lot better going forward for your visualization needs. And then, as we wind up here with these tips I'm looking at, the next one I have on my list is a function called rel Now where this comes to play is when you're sizing things such as, you know, the base size of the font you have going. In an example here, she's got a facet or a box plot

with the country on or the continent on the y axis, the population on the x axis, and the plot has a title at the top, and the base size that's being used here has been set at 13, then the plot title was size of 16. Now what might happen is you might have an issue where in the first glance there's something very wrong with this, but when you resize the plot for any reason,

then you've got to increase or decrease the font size, and that's where you've got to figure out then, okay, what value should I put for the size here and there? Another manual kind of trial and error effort. But there is a better way because if you want to set it and forget a little bit, when you do the plot title in this case, you can instead do a call to elementtext where you give it one of the parameters

for the theme itself is size, but then instead of just putting a number for the size, you can surround that with the rel function, which means this size is specified a relative font size based on the size of your plot itself, and in this case, this means that the size of 1.25 that's supplied in the rel function is simply meaning that the font will be 1.25 times larger than the base size,

and that will be consistent no matter what you do with the rest of the plot function. I use rel all the time for a lot of these cases because this, this old guy here doesn't have time to troubleshoot font sizes with the default numbers, number, invocation. So having REL is a lifesaver when you want to keep those proportions, but you're not quite done yet tweaking what the size of that final plot will be, the resolution of it, and other considerations. So that's another nugget that you want to put in your toolbox

next time you're dealing with these different sizes and you just want the font size to just work. I think this is a handy, handy feature to opt into. But we have much more than just that, Mike. Why don't you take us through the rest of the post here? Oh, I wish I knew about RHEL before this. That's a huge tip. Yes. But all of these are underappreciated, great, really practical tips. The next one is the plot dot title dot position argument in the theme.

So I think by default, and Nicole, demonstrates this with a bar chart where the coordinates are flipped essentially so that you have your labels along the y axis, and then you have kind of your continuous numeric values along the x axis. And if you stuck a title on that chart, you know, sort of just by default, that title would be pushed into, kind of above the bars and not actually above the the y axis labels.

And one way to fix that is to use within the g g plot theme function, this plot dot title dot position and set that equal to, plot in quotes. And that will actually push your plot title all the way to the left of your visualization and stick it, you know, right over the top of those y axis. Values looks cleaner, more professional, and you have a lot more space in case you have a long title or a long subtitle or something like that. And you can essentially, you know, plot dot title dot position, adjust both the title and subtitle. If you have a caption and you're looking to do the same thing, there is a plot dot caption dot position argument that works just the same. So that's a great nugget. The next one is

the b g argument in g g save. And sometimes, depending on the theme that you're using for your g g plot, the background of that plot may actually be transparent. And when you go to save that, sometimes transparency can be good. Right? If we're trying to stick this slide in a slide deck that already has, like, a background theme that we want to match to. But sometimes it can be bad, and we actually want to enforce a background in our, you know, dot PNG or dot JPEG

object that we are saving locally. So you can use within the g g save function an argument called BG and set the background to a color that you want, you know, such as white. So that can be really handy. And there is one last sort of bonus function that Nicola points out here, and it's scale x or scale y underscore symmetric. And this one is really cool too. It's not actually from g g plot two. It's from the lemon package, another one I was not familiar with. But it essentially allows you to create,

a symmetric axis, either your x axis or your y axis. You know, typically, you know, for example, around zero. So let's say that I'm plotting residuals, you know, from my machine learning model, and I want to see the distribution of those residuals above or below zero on the y axis. Sometimes if some of the more of those residuals fall either below or above the y axis, zero might actually not be in the middle of your y axis. And it can be a little misleading for folks who take a quick look at that residuals plot and just sort of assume that zero is right in the middle and then see dots above and see dots below the halfway point of that y axis and think, oh, everything's evenly distributed.

Well, if zero is much further up on that y axis or much lower down on the y axis, then that is definitely going to mislead them. And if you want to adjust that and sort of force zero to be in the middle of that axis, that's where you can use, you know, the scale y symmetric or scale x symmetric depending on which axis you're looking to adjust, function from that lemon package, and it will do that for you. So really handy. Oh, wow. I that would have been so much useful many years ago when I was doing these, biomarker

kind of regression plots and we were interrogating just how well the fit was and these residual plots I was doing were pretty misleading. I had to customize the axis, you know, tweak marks on my own, but this symmetry symmetric function all in one shot. Wow. Yeah, just goes to show you that it's not just ggpod2, but the ecosystem around it. And heck, I always like a package that's named after a fruit, you know, can't just have apple be everywhere. What's good they offer fruits a

fair shake too. Yeah, I love these kind of posts because there are some gems in here that you can work around yourself, but you're going to save yourself a boatload of time by having these, in your toolbox or at least documented somewhere in your notes so that you, or if you're lying on a friendly LLM provider, they can also use these use these great tips as well. Totally agree. And machine learning has definitely stolen all of the vegetables in terms of carrot and parsnip. It's about time about time we start using the fruits.

And last but certainly not least, then the highlights today. Have a great blog post from a good friend of ours in the community that kind of risen back and is blogging, you know, very regularly now and not stopping maintaining some very core infrastructure in the world of R itself. And we're talking about the esteemed Yvesiye, of course, author of so many great innovations in the R community. And we're going to talk about some new features

that are landing in arguably one of his most influential contributions, and that is the Knitter Package. And just what is landing in a knitter package that we're going to talk about today? It is actually support for the newish types language. And what is types, you you may say? I mean, not so much language, but, document typesetting. Types has become a very interesting, you might say, alternative to what has become ubiquitous

in this world of LaTeX for document type setting, and yes, TypesT is written in Rust for a lot of reasons I know of, but one of them is the performance and having also a modern approach to bring some semblances of web styling into static documents. Well, as such, types much like LaTeX before it, have its own DSL. So in the upcoming releases of Knitter,

or actually maybe already out now, it now supports a new document type with the extension .Rtyp, much like if you remember the early days of Knitter, it has support for the sSweep or Sweep syntax with the capital R and then lowercase nw, and of course capital RMD for R Markdown that most people use it for. Well, now R type is now joining the party here.

So the blog post starts with a quick example of a very, you know, rudimentary hello world type of types document, where you kind of see the type syntax at the top with the page size and or the yeah the page layout and also the text you know font and size with that in a a title along with now the very familiar r code chunk which does not have any change from what you've done in R Markdown or R and W before that, just the typical fenced, you know, callouts with a bracket R closing bracket or squiggly bracket, I call it, and you can compile that with Knitter, and there you go, you got yourself

a new PDF as well as the type document itself with extension TYP if you want to look at that alongside the PDF. Again, just like how the rest of the R Markdown and LaTeX formats work in the knitter engine itself. Now we get to talk about some of the, you know, more humorous parts of this post, but also they ring true for a lot of reasons. Ewei has been very upfront

with his advocacy, much like how I've been trying to do, of the HTML format in as many situations as possible for a much more richer interactive experience, And yes, the con has always been over the years, well, you got to print it out for some people and sometimes HTML doesn't print well. Well, sure, types is going to be a great alternative if you need the static type document that can look good when printing. But anyway, he's also been

delving into space on the HTML side, and what comes to mind is a package that he created totally on his own many years ago called Pagedown, which was a front end to the page. Js JavaScript utility, which he built himself. It's a literal page. Js script, and if you use, say, Lightdown, which is his newer R package, as kind of a minimized,

you might say, less dependency footprint version of what he did with R Markdown before that, it will use this PageJS under the hood for a printed layout, and from his testing, that's looking pretty solid. Now I know, I'm sure Mike will comment on this too, sometimes we don't get as much choice as we would like on the formats. Yes, I know, I know, but if you are confined to Types or LaTeX or any of the Office formats,

at least in R, you now have, in the case of Types, yet another way to opt into that, even if not you don't really want to necessarily opt into the Quarto, you know, ecosystem for type support. Now you can have it in Knitter itself if you prefer that layout as well. So it's there, he is committed to making sure that's mature. He acknowledges he will not be a user of this feature because, again, his world's in HTML for what he desires

the rest of report generation to be. I'm with the OUA, I want HTML to be a first class citizen, one thing at a time, of course. So yeah, there will be many people to thank you for NINR support and types as we navigate waters and this modern way of disclosing and communicating our data science results. Always great to have E Way in the highlights again. Mike, what did you think about the type support here? I am so jealous of E Way having the luxury of not needing to use,

you know, latex PDF types type of utilities to generate PDFs for clients. Man, we have spent so much time on wrangling the dependencies, you know, and put a lot of work open source as well to try to help people with that sort of exact problem. And maybe I should've just been sharing his JavaScript script the the whole time. If I get my hands on that, that would be pretty cool to see. But I think this is

really helpful. I I have to imagine that, you know, this sort of bridge that Eway has built between R and, you know, types style, extension files is, you know, not only gonna be useful for this sort of direct use case that he lays out, but potentially for the future integration between R, Quarto, and Types, which I I feel is only growing. You know, we have migrated completely to a a Types shop. We've gotten away from LaTeX and the the speed is sort of unbeatable.

You know, talks a little bit about the bottleneck of, you know, while Types is fast, the computation in code chunks can be slow in all of these files, whether you're using LaTeX or TypesT. I think, you know, we have laid out sort of a a path, you know, that I've shared in the most recent, you know, risk conference for the R Consortium where you're using targets initially sort of to build all of your analysis ahead of time and your plots and your visualizations and tables and things like that and pre computations.

And then once you've done that, you you have this sort of DAG that the only the things that have changed need to be rebuilt. And when you go to render your types or or R Markdown file, you know, all of those computations are already taken care of. So it's it's lightning quick. And at that point, when you're comparing LaTeX to types, I think types really flies. And especially when you're working on a big project or a big document, that speed of iteration can really build up over time and save you, you know, minutes, hours by the the time that project is over. So that's a little rant on sort of my support for

types, but very grateful that is is doing this type of work and has laid out this really nice example for us. And I certainly wish that the whole world could just go HTML, and PDF could be a thing of the past. Yeah. Or even more of a hot take. Please, for the love of you know what, stop doing PowerPoint all the time. Oh, gosh. There are people at my org that are making clawed skills

to make better PowerPoints and I'm like, no, stop. So we're trying, we're fighting the good fight, but it takes a while. Two steps forward, three steps backwards. Yep. Yep. And the officer pack should get you kind of far, but you're still, yeah, missing out on the good stuff. But where we can control it, I am advocating for at the minimum slides in HTML with Quartal, and obviously I'll be keeping an eye on types for the PDF generation if we have to keep in

that ecosystem. By the way, Mike, is the recording for your Our Consortium Talk already available online? Yes. It is. Do know? Oh, yeah. We'll put a light in the show notes as well. It was a great talk, and everybody should watch that. You are you've been ordered to do so by your host here. But, you know, what I also would like to order or at least can and gently encourage you to do is, of course, visit the rest of the issue at rwika.org because these three highlights are just the tip of the goodness, so to speak, of what's available in each issue.

And we can take a couple seconds or minutes here for some additional finds here. And so what I want to shout out is an interesting post, actually it's a GitHub readme, from the esteemed Dirk Adebutel that talks about, with the recent release of R version 4.6.0, what is happening with binary package installations? And that is, of course, something that's very prevalent, or very important, I should say,

for those of us on the Linux side of compiling our packages. And so he's got some nice treatment here about best practices for detecting whether packages need to be rebuilt into their new binary format and looking at ways of having to deal with that and the implications that might have to downstream effects on this, such as the graphics engine and the different results across different Linux distributions.

So the post has some good call outs to the different R packages that where binary versions are most definitely helpful with and looking at the different trade offs that are being done to update these, but if you're into the Linux space with building packages, I think DirXposed might be worth a read if you're upgrading your version of R and seeing what the impact might be for compilation and that binary generation. So as always we'll put that link in the in the show notes as well.

So there is a giant blog post here titled exploring and learning quantum computing part one, I can't believe it's only a part one, by Ken Kunwong. And if you have any interest in quantum computing, and a lot of this is slightly over my head, but it's phenomenally laid out, from what I can tell. It's leveraging, you know, both r and I think a little bit of Python too via reticulate to show some quantum computing, examples. I think it's a a fantastic one to check out, so I'm glad that one made the highlights.

Yeah. I know that's starting to get a little more attention as the, obviously, the hardware architecture side of things with the advent of AI and data centers spinning up almost everywhere else is, they're thinking about other ways to leverage that computing. Quantum computing, I think, is going to be some of the talk about in the coming years, but like you, Mike, I wouldn't even know where to begin. So I'm interested to see where this post goes, but it's a good accessible

treatment of this because I know there's some very enthusiastic members out there that want to be on the cutting edge of this, and this is a great way for us to do this from the friendly confines of R itself, a little Python help here and there. So great, great find there. And as we mentioned, there's a lot of other additional content in the issue, and we invite you to check that out. At rweekly.org. As always, the home page always links to the latest issue, all the sections that you can find for new packages, updated packages,

even some fun images with artistry and whatnot, and I had fun putting those in the post today because even our our friend, Mike FC aka Cool But Useless is back at it with turning R into this retro gaming engine that will make me have no free time whatsoever. I digress, but it is, a great issue, through and through. And we of course depend on all of you for helping us find these great resources on top of our RSS feed mechanism, which if you have a site and you want us to syndicate your

RSS feed for potential inclusion into the issue, have we a way to submit your site's RSS feed, but also if you see a specific post that you want to highlight or at least get it into the issue. We're just a poll request away, folks. Go to rwgo.org. At the top right corner, you'll see that little GitHub ribbon of sorts, and you'll be taken right to the issue draft for the upcoming issue. You'll be able to fill out a simple template for the PR, and we'll get that, you know, merged in, and we'll be glad to have your contribution. You will be helping the project by the community and for the community.

And speaking of community, we like to get involved with the community as well and see you know dialogue with all of you. Where can you find me? I am on blue sky with rpodcastbsky.social. I'm also on Mastodon where I'm rpodcastspodcastindex.social and I'm on LinkedIn. Just search my fancy name and you'll find me there causing all sorts of chaos or whatnot or even just lurking here and there. Mike where can the listeners find you? You can find me on blue sky at mike thomas.

Bsky. Social or you can find me on LinkedIn probably not if you search my name because there's too many of us but if you search for Ketchbrooke Analytics, k e t c h b r o o k, you can see what I'm up to. Very good. And, again, we'll make sure in the show notes we link to your excellent talk from the risk conference because that was that was a great, great session there. Speaking of conferences, I do have a section in our weekly about that as well and there's also ones I may not even have mentioned there such as the upcoming

Gen AI Our Pharma Day in June where we're going to talk about some great use cases of AI in the world of life sciences and other events that are happening. Nonetheless, we're going to close-up shop here. We're going to get this episode in the books. Thank you again so much for joining us on this week's R Weekly Highlights episode. Very happy to have you listen from wherever you are and yes we will be back with another edition of our weekly highlights next week

Open the transcript file

Hosts and Guests

Hosts

  • Eric Nantz Host Feed-stated
  • Mike Thomas Host Feed-stated

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Issue 2026-W47 Highlights

    Nov 20, 202648 minAudio

    Our takes on the increasing removal of friction in good and not-so-good ways thanks to AI, a visual tour of hot air balloon accident data, and a case study of novel image recognition with the rplatipus package. Episode...

    Credit the people on this episode

  2. Issue 2026-W46 Highlights

    Nov 13, 202633 minAudio

    The flow of both data and analytical jobs just got smoother with mizu, opening the door for multiple agent harnesses in Positron, and an optimal statistical model tailored to complex experimental designs with a friendly...

    Credit the people on this episode

  3. Issue 2026-W45 Highlights

    Nov 6, 202644 minAudio

    What happens when that online data set isn't so easy to download anymore? We learn of two packages to ease the pain of obtaining these files and then some. Plus a visit to the visualization corner with a handy way to...

    Credit the people on this episode

  4. Issue 2026-W44 Highlights

    Oct 30, 202631 minAudio

    The celebration of rOpenSci's 15-year anniversary continues with the community sharing their perspectives, and a fun Shiny app that could be quite useful during celebrations on a weekend or the holidays. If you ever...

    Credit the people on this episode

  5. Issue 2026-W43 Highlights

    Oct 23, 202659 minAudio

    We learn about the variety of ways you can contribute back to the core of R, a roundup of the excellent choices available for package and environment reproducibility, and an amazing collection of Shiny applications and...

    Credit the people on this episode