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-W35 Highlights

Aug 28, 202641 minAudio
ShareXLinkedInFacebookWhatsAppEmail

A collection of techniques to unlock creativity for your next visualisation, a look at the exciting present and future of the futurize transpiler package, and how to avoid an all-to-common pitfall managing database connections in mirai daemons. Episode Links This week's curator: Eric Nantz: @rpodcast@podcastindex.social (Mastodon) & @rpodcast.bsky.social (BlueSky) & @theRcast (X/Twitter) Why creativity matters in data visualisation useR! 2026: Futurize - Tearing Down Parallelization Barriers in R with Transpilers Tricky Parts of Async Mirai Magic Entire issue available at rweekly.org/2026-W35 Supplement Resources When creativity meets code slides and abstract https://nrennie.rbind.io/talks/ross-ihaka-lecture/ When creativity meets code recording https://www.youtube.com/watch?v=fAiyiE8Pm8s {futurize} documentation https://futurize.futureverse.org UseR! 2026 {futurize} slides https://henrikbengtsson.github.io/talk-user2026-futurize/ {mirai} documentation https://mirai.r-lib.org 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 Warriors of Light - Final Fantasy - Darangen - https://ocremix.org/remix/OCR01339 Ice Cap Zone (Pulse Mix) - Sonic the Hedgehog 3 - MkVaff - https://ocremix.org/remix/OCR04400

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. This is episode 239 of the Weekly Highlights podcast. This is the usually weekly show where we talk about the great resources that have been shared in this week's issue of R Weekly, which you can find at rweekly.org. My name is Eric Nance, and I'm delighted you to join us from wherever you are around the world. Yes. We were off last week, only well, there are a couple of reasons, but one of big ones was we unfortunately did not have an issue last week. But this week, we do have an issue, and it is actually a supersized one because

we combine what would have been in the last week into this week's issue. And normally, I am not alone on this, but unfortunately, this week I am alone on this episode. Mike is, tied up with current business travel and he can't make it today, but rest assured we'll do our best to get him back on the actual microphone here in next week's episode. But for now, you're stuck with me, your trusty host Eric Nance here, and again, very happy to join me. You know, even without Mike, we're gonna do our best to have a have a fun episode.

What do you think, mister MCP? There's a 68.71% chance you're right. Those odds sound pretty legit. In any event, we're gonna have a fun set of highlights to talk about here. And, as far as wearing multiple hats, well, guess what? This week it just happened to be my week for the curation, so I am pretty well familiar with the highlights here because I helped set up the poll to eventually pick them. But as always, I have tremendous help from our fellow RWKY team members both in their contributions,

feedback, their votes in the poll, everything all in between, and contributors like all of you around the world with your poll requests and other wonderful suggestions. You know, especially in the recent years, I would say it can be pretty difficult to get attention when you need it. No, I'm not talking about, you know, other things in life like when your kids want attention in parenting, that's a whole other topic altogether. I am talking about when you have a message to convey

to make sure that the audience you're trying to target is really understanding and resonating with the message you're trying to tell. And this is very true for us in data science, where a lot of the ways that we communicate our analytical results, our key findings, is typically through a visualization in a graph. Could be a static graph, could be an interactive graph, but that is usually how we tell the story, so to speak, of the data that comprises that result we're trying to demonstrate.

And, you know, there are a lot of times strict requirements on these visualizations when you get to more of a rigorous, say, manuscript, you know, process, or in my industry in particular, some of the do's and don'ts when it comes to visualizing clinical data, but there are other cases where you are trying to, again, maybe convince an audience of what is this result really useful for and maybe getting a call to action as a result of it.

Therefore, it's probably not just quote unquote good enough to give a technically valid, valid, visualization, but you need to up the ante a little bit and maybe show us some of your creative juices to get that visualization's message across in ways that your audience can really resonate with. And these principles come from a very thought provoking blog post with an accompanying lecture, authored by the esteemed Nicole O'Reny, who has been featured in highlights. I probably lost count how many times her blog posts have been highlighted here on the show,

but she comes to us with this latest blog post entitled Why Creativity Matters in Data Visualization, And as I mentioned, this is actually very related to a talk that she gave at the very highly regarded Yihaka lectures at the University of Auckland, where our favorite programming language on the show, the R language, was originally born. And she gave this talk at the end of July and you can actually view a recording of this talk, which I have linked to in the episode show notes if you want to see it, see the talk for yourself.

And her main ideas here are that it's not just a good visualization, so to speak, if it's technically correct, but a lot of times, again, depending on the message you want to convey, there are other qualities besides the technical bits that you probably want to get a good handle on if you want to convey that message, and she breaks it down into three main aspects, and she admits that some of these can have a negative connotation,

but in the lens that she shows here of examples, I think they definitely resonate with me. One is grabbing the attention. Like I said, attention's at a premium these days. Everybody's so busy, everything's competing for attention. So you don't want to just read a wall of text instead of a visualization, that's one thing. But maybe there's a new way to present the aesthetics of that visualization in a way that really relates to the data itself that might be helpful. Another quality is whether

it is memorable, that when somebody takes a step back after seeing it for the first time, does it stick in their mind? You don't want them just to notice a chart just in a first glance, but also the message behind it, and sometimes the defaults that we get in our popular visualization packages may not really lend itself to that, you know, kind of sticking into someone's memory. And lastly, it can have some emotional ties to it too.

You want your audience not just to see something, but if it's really to spur on a call to action or maybe, you know, another result or a decision, you want them to really feel that, so to speak. Now, this isn't always just an artistic thing, and of course, you could really muddy this up when you talk about politics as well. I'm not getting into that here, but there may be kind of an advocacy

side to this where you want to really get that audience engaged for next steps, And that can, again, mean a lot of things here. And what Nikola mentions here is that a lot of these ideas actually have been in practice for a long time, even in some of the historical visualizations. So it's not just about trying to be quote unquote modern, but these approaches have been demonstrated in history, and she has some references

to this as well throughout the post that you can see for yourself. But that's kind of the three key qualities, but of course, you as an R user, a data scientist that's creating a visualization, what are some techniques that you can use to get yourself into that mindset, even if it's not such a comfortable mindset for you. I'm definitely not a creative person in terms of visualization, although there are some times I feel like there is an idea I want to convey,

maybe based on a time of year or something like that, but these techniques here I'm definitely going to take a lot to learn from, one of which is you yourself let the data inspire you and letting that data help dictate what type of chart you want to display. We're not just talking about, oh, for categorical type data, do a bar chart, and for quantitative, continuous data, maybe a line chart over time, you know, things like that. We're talking about

another aspect of this that really takes into account what the data actually represents. And so she has this wonderful example of surface water in Great Britain, but the plot kind of looks like a map of England, but the different regions in terms of the prevalence of the water and surface water is represented by bubbles, literal bubbles, not like a bubble chart. I'm talking about bubbles that you might see in

water or whatever. And the bigger the bubble, the bigger the surface area of water. And that's really it. It's actually a very monochromy looking plot, but it's minimalist in all the right ways. You will remember this. What other chart have you seen that quantifies, like, the density or frequency of something by an actual physical thing that can be resized. That is really interesting. So it took the idea of a map,

and instead of doing a gradient of colors to show intensity, it did something that catches your eye, and you can see in this map the northern regions of Great Britain had more of the water surface area. So it's very easy, gets your attention, and a novel way to let the data dictate what you're trying to show. And then up next, it's not just the type of the chart, maybe the data itself lends to what you actually choose for some of the aesthetics in your plot, such as the color choices.

You know very well, if you've been using, like, ggplot2 in the R ecosystem, you get that default color palette, you've seen it in all the online textbooks or the printed textbooks, you know what they are. That may not be appropriate for the type of data you're looking at, and so the next example that she shows off here is a line chart showing the baking ingredients that were used in a television series called The Great British Bake Off. This was, produced by Kara Thompson.

And guess what the color choices represent? Well, one of the key ingredients is chocolate, is one of them, so that's got the natural brown for chocolate, but then also orange for the orange and raspberry, a reddish color for raspberries. Makes perfect sense. They come from the actual data itself. That is something that you can use as well just to give it that extra bit of attention grabbing

that can really help in these situations. So that is also very, very helpful, and you also might want to, you know, keep in mind the color choices may not always be the most accessible in that situation. She recommends looking at, assessing accessibility in your visualizations along with linking to Kara Thompson's previous talk at this at a twenty twenty five RSS conference, so you'll want to take a look at that.

Last but not least, you can also let the data choose what kind of style you want for your visualization. This one, now I'm going have a lot of fun with this because it's right up my alley, where, let's say you're looking at a domain that's a little bit more technical, might say retro technical, and this isn't just talking about in the world of text, font choice is often used to, you know, let the style dictate what you're doing,

But the chart that she has, an example here, I'll tell you what it's about, and then I'll tell you about the choice of aesthetics here. Let's talk about programming languages from a programming language database and how much prevalent they have been in the different rankings. And what you use for this is actually straight out of what looks like a a retro terminal from the Commodore 64 days,

a monochrome black background with green text and computer looking font. So she used this for looking at the popularity of languages listed in the programming language database and what kind of, symbols they use for different comments and the different, tokens, like such as a hash sign, two slashes or backslashes, or even a punctuation symbol.

And you see, depending on the language, you know, much different rankings there, where obviously Java has the two backslashes, whereas Python has the hashtags along with R itself as a common choice. But that was, again, you could have done this in a plain aesthetic, but instead it's a computer techie type of style. I am here for it, especially as someone who literally authored about a month or so ago a chordal theme that's a retro terminal theme.

I could definitely get behind a visualization like that. So those are techniques or principles that you can follow, but it's not just that. It's about when you're actually scoping this out. You know, I'm guilty as charged about just turning loose in R itself with my favorite R package and just hacking together some visualization starting from scratch, add layer and layer, but I'm in the confines of the framework itself. She challenges you, the audience, are reading this or hearing about this

that maybe go to the classic pen and paper or some interactive digital way of drawing this before you actually get to code, and it kind of unlocks your brain a little bit of what you can create. So she's got an example of a chart that, you know, this technique was used with that she made about the prices of peach jam or the percentage of this in regards to tariff rates

for the different types of jam, not just peach, but other types of jam out there that you can get at a grocery store, and she put each of the different types of jam in a vector based jar with the level of the jam corresponding to the percent of the tariff rates. How cool is that? That is really slick. And she and again, she started hand drawing it and what this jar might look like, but with a little investment in time, maybe of a little help from your favorite, you know, large language model,

you could build these type of visualizations even in code, but you may not think to do that if you start with a code or that, you know, interactive visualization editor, whatever you prefer. That is a really interesting idea here. So I I mean, my pen and paper drawings are terrible, but if I ever am in this space trying to communicate something that isn't in such a regulatory mindset, that might be a technique I have to, you know, invest in more over time.

There are a lot of other great ideas here, such as don't add more just for the sake of adding more to your visualization. You may just wanna do a minimalist approach to conveying that information. She's got a lot of lot of ideas that have drawn inspiration from other creative endeavors as well. So there is a lot to digest here. I have not been in this space of maybe occasionally trying to convince

an audience based on a visualization I've produced, but I can see situations in the future I might have to be that way, especially if I have to advocate for more investment into, say, a project or an application I'm building and trying to convey more than just how many people are using my package or how many times my application was used, but try to maybe convey that in a more eye popping manner. So I will file this as usual in my massive collection of bookmarks,

but Nikola, always, always one of the most prominent thought leaders in this space. And the talk, the Rasa Hakka lecture that she gave in July, highly recommend to take a look at when you have a chance.

And up next in the highlights, we're gonna you're gonna speed into the present and the future quite literally with one of the amazing efforts that has continued to evolve, which has now a probably close to five year or so history to bring much needed convenience with high performance computing in the R ecosystem. I am which ecosystem am I speaking of in particular in this? That is the future verse.

We have a set of packages that have been originated by Henrik Benson, professor at the University of California San Francisco and one of the original innovators in the world of R and HPC. I have so much to thank him for. I still remember fondly having dinner with him at the end of one of the posit comps and just kind of being that sponge, soaking all the knowledge of his past experience when I was head deep in HPC computing back in the day.

Well, what are we talking about here? This is actually a recap of what he talked about at the recent USAR twenty twenty six conference about, in particular, the new package entry in the future verse called Futurize, and the talk that he gave at useruseR was called Tearing Down Parallelization Barriers in R with Transplylers, and how it's easier than ever to do this.

This is massive and a lot of ideas I think we can take in any ecosystem that involves, you know, a situation where you have to go from the default behavior over to a more performant behavior without having to refactor everything. So let's motivate this a little bit, and we will have a link to, Hemric's slides in the show notes as well, so if you want to take a look at the slides of that. The recording of the talk is not available yet, but I'll keep on the lookout,

from the user organizers on when that takes place. Historically well, easy for me to say. Historically speaking, when you have tried to do heavy demanding computation task, one of which could be bootstrapping and generating a bunch of replicate datasets to construct your confidence intervals for a bootstrap, technique. If you're doing that all sequentially,

sit back, relax, get your favorite beverage, and it's gonna be a while. It can take hours, sometimes days. Same thing for analytical modeling, such as mixed models, where you have to look at different subgroups of patients and fit a different structure for each of them. Good luck. Have some patience. And one that really resonates with me, building simulated data to design your experiment. Yeah. Do you want to wait overnight for 10,000

iterations to finish up? Or do what I did in grad school, SSH to your Computation Labs Linux cluster, if you want to call it a cluster, launch that R process in a TMX section, disconnect, and hope for the best when you wake up that it worked. Yeah. That's what I did. I know many of you have too back in the day. Well, if you're doing that sequentially, your little computer there is not doing very much if you look at the actual,

you know, real time metrics such as utilization of CPU cores. But if you can take advantage of those multiple cores or multiple nodes in a cluster, you paralyze everything. That's what you want to do, embarrassingly parallel domain. Now in the past, if you wanted to go from that sequential to the parallel workflow, you had to refactor. A lot of times not by choice, you had to do it, and it could cause a lot of friction.

What function do you use? Is it from the parallel package? Do you have to make your cluster things ahead of time? Whereas you had an l apply call in the sequential sense, you've got a huge, you know oh, not huge, but probably like a ten, fifteen line set of code to deal with a new version of l apply. Well, there's a tax you have to pay for that. But what if you can have your cake and eat it too? You start with that original sequential code, but you can just tack on something

to make it more parallel without much effort, and that, to get to the punchline here, is where the futurized package comes in, where instead of having to even use one of the future related packages to do its function code, you can now feed that original sequential call like lapply into literally the futurize function. What that futurize function is doing under the hood, it is translating

that original code, the l apply, on the fly to what would be the future versus version of that. So, for example, the l apply call. When you pipe add the future eyes, you don't see it in your, in your immediate session, but behind the scenes, that has magically turned that function into the future lapply function from the future dot apply package. You could have written that yourself. If you're an expert in the future verse, you probably are comfortable.

But if you're not, the futurize function pays off big time for making that work, and it is very low friction because it applies to not just the classic base R functions, it applies to other ones as well, whether it's from the it's a map function in base R, whether it's from the per package. You can also use this not just within nodes, but like by Rest of the Futureverse. You can go from utilizing a single machine very well to using cloud computing

and HPC with the different integrations you can have for HPC clusters, for Mirai processes in general. So many things you can choose from there. And all you have to do is just set up the plan ahead of time, and then you can still use Futurize to respect that plan, and you're off to the races. And as I said, it's growing in terms of what Futurize is capable of or compatible with

in the R ecosystem. Whenever you're doing statistical modeling and machine learning with the carrot package or the party kit or any of these other ones in the world of genetics and genomics with Bioconductor. New packages are becoming compatible with Futurize as well there, too, and I think this list is just going to get bigger and bigger. I mean, this on the simulation side, this is going to, you know, really speed up things

when you can distribute all this in parallel workers and use that same code to do the bootstrap, literally to function boot, and just feed it to Futurize. It's a beautiful thing. You want to see how long something takes? Guess what? You get that too because you can use this with the progressify function from the progressify package to feed that in your pipe before you feed the futurize. So it'll go, say, your l apply call, pipe the progressify,

then pipe the futurize, and you get that nice command line progress bar with the ETA and the percentage complete. Really, really cool stuff. Again, two lines of code, and you could keep the original code unchanged. There is a lot coming in the future of future eyes and the future verse in general, some of the ideas that Henrik, talks about here, the idea if you can cancel a current, call,

if there is an error or some kind of interrupt from the user, Also looking at whether parallelization is even worth it, getting an estimate of what efficiency you're getting or maybe even not getting. Also taking advantage of new developments from the Charlie Gao, for instance, at POSIT, one of the software engineers there, with a package we've highlighted recently called Mori for memory and distribution of data, you know, optimization

in parallel workers, Futureverse wants to be compatible with that, as well as having a similar syntax as this Futurize call to declare resources in a PIPE friendly fashion instead of like that plan step where you have to opt into one of the future verse kind of back ends. Instead, you could just have a call to resources, give it what kind of memory you need, what kind of CPU or GPU cores you need or how many you need,

and the time that you're willing to wait. Like, what is the limit or the max that it should complete? All that is in the works, and it is all very fascinating to see. Again, all the packages are on CRAN and on GitHub. The slides have links to everything, so I'm really excited to see the future verse evolve even further. It's also got some of the best logos I've seen in the R ecosystem because I am a big fan of Back to the Future and you can definitely see the influence there and the typography

in the hex stickers. So Henrik, definitely if I see you in a conference in the future, I'll shake your hand again for thanking you for all the great developments here, and I'm very excited to see what is in the future for the future verse. And last but not least in today's episode of the highlights, we are gonna stay with kind of efficient

high performance computing and now going to another popular framework that we've been seeing take a lot of foothold in the R community and that is the Mirai package and some of the common trickier parts you might have to think about if you are launching these Mirai processes with other integrations in your data processing flow, in particular database connections. And this is a continuation

of a series that has been featured recently in the highlights by the esteemed Albert Rapp as part of his three minute Wednesdays, series of posts, and that is where we're going to talk about a case study of when out of the box, you might be left scratching your head a little bit of why everything works fine locally, but then with MirrorEye, suddenly things start to break down without a little bit of help along the way in your database adventures.

So what is the problem we're trying to solve here? Well, in the R ecosystem, we have packages that are interfacing with databases that leverage the DBI family. And then within that, you'll have domain specific packages that leverage DBI on the back end but give it the necessary specifics that talk to that particular database. We have connectors as SQLite, which is demonstrated in this post, Postgres databases.

You know, DuckDB is one of them too. You got a lot to choose from here. So let's say you've got a SQLite version of a database of very fond PalmerPenguins dataset. You can establish that connection very easily with the DBI colon colon db connect function call. Within that, you feed in that more domain specific call to R SQLite to establish that connection with the file path. You've got it, everything's there, you can query that data, you can, you know, transform it if you need to, all that fun stuff.

You've got now that connection, and you want to start thinking about using that connection in MirrorEye processes, where with MirrorEye, you can simply call the MirrorEye function, feed in what you want to run, and then if that call what you want to run depends on an object, you feed in the reference of that object from your environment, in this case the con object for the connection, and when you run that out of the box, you get an error saying it's unable to find an inherited method for function

DB list tables for signature con equal SQLite connection. So you're wondering what the heck's going on here. It worked just fine outside of MirrorEye. What happened here? Guess what? That DBI call, which was the initial call to DB Connect, it comes from the DBI package, that was explicitly defined, but the Mirai process does not know about the other package that was involved in that, and that is R SQL lite.

It is not actually loaded in the daemon worker that was spun up. So seems like, okay. Well, remember what Albert told us in the last time he did this is that in Mirai, there is the function called everywhere where you can load in packages that should be referenced in your session. So you're thinking, great, I'll call Everywhere, feed in both a library called the DBI and R SQLite, and think, okay, yeah, that'll take care of it. And kind of good news, it does

help with that particular error, but now you get a different one. The error now says bad weak PTR. I don't know. I know a lot of weak things. I don't know what a weak PTR is. Well, that actually stands for pointer. This is niche in the concept of database connections, but it is important to figure, to keep in mind. When you use a DBI compliant, you know, way of connecting to databases,

that connection object, if you actually look at what's in it, and he does an example here of running the str or str function on it, you'll see in there it's got an attribute called pointer or ptr, and it references something called external ptr. That external, not external outside your system, it's external to the R process because it is pointing to an actual memory address in your system.

There's a lot of nuances there, and technically, R objects do have an address as well, but databases have a more special treatment of that. And basically the Mirai Daemon process can't get to that address. It's like there's a roadblock on the way there in your system when you have the Mirai processes. Okay. So now what do we do? Well, maybe what you can do is simply call that dbconnect call within

the mirai worker itself and put that as part of something like MirrorEye map, where you just run this three times in this example, and you're establishing the connection right there, and he shows what happens when you are trying to do this new connection per worker instead of just that default connection in your global environment. What's interesting is that you will now not have an error anymore,

but you will notice each of these workers is getting a different address for this pointer from each other. So in the example he shows here, there are these random ish, notes, zero, x, and a bunch of letters and numbers, but they're all different. You may not want that depending on what you're trying to do with this database, especially if you don't want to flood it with a lot of connections at the exact same time.

SQLite in particular doesn't really like that. Maybe our Postgres likes it, but not usually SQLite. Now what if you want to get back to what you're originally trying to do? You have one connection, share it with all the workers in the most least friction way possible. Well, one other way you can do this is in that Everywhere call, you don't even necessarily have to load the packages like we did in that previous example.

You can use the operator that I've always been a little scared about. I just had bad experience of this in the past, but you can do this. How to say this in words, the global assignment operator, which is basically the two less than signs and the hyphen. So the typical assignment operator is the less than and the hyphen. The global assignment operator is the two less than signs and the hyphen. That means you are assigning this variable outside

of where you're actually calling that variable in to be a level up, so to speak. So in the case of this everywhere function and you feed in the con, you know, the liner for the DB Connect to assign to the con object, that con object is actually being assigned outside of everywhere. I admit, even I'm kind of fumbling my way through that explanation, might have to play with this a bit. But it turns out if you do that and then have that mirai map call that references the con

and you want to do, like, a simple list in your tables there along with that memory address that he showed, you will see that the code works, and you get that same memory address for each of the workers. So, again, this may be very helpful in the SQLite situation, but your mileage may vary. Maybe you wanted that separate connection per worker. What Albert shows you is that there is a way to do both. You just have to do it a little differently, but you can do both.

So you gotta explore that for yourself, but it's nice to know with mirai, you just have to have a little bit of proactive thinking here to make sure all that works seamlessly with what you're trying to accomplish. So as always, Albert does a wonderful job of distilling these concepts. Yes, they are a niche of a niche per se, but when you need it, you really need it and you need to understand how it works. Albert's post here is a great introduction

to how you might have to handle this, whether it's database connections or other parts in the R ecosystem where you're not just getting an object in the traditional sense, you're getting a pointer to something else. Well, let me point you, so to speak, to the rest of the issue of R Weekly. Like I said, this is a supersize issue. A lot a lot of great posts here because they cover two weeks of of an issue here. So as I'm scanning through this, I don't have time to go into each of these in-depth,

but there were some interesting blog posts such as from Jumping Rivers, looking at the trends of what's happened this summer in the weather standpoint, looking at some new frameworks for reproducibility with the UVR package and how that applies to scripts and quartal documents, very interesting, as well as a new project from Henrik Benson himself about trying to get R to tell you proactive or how you can tell R how much memory you need. That's a fun project that's just getting off the ground. Lots more to choose from there, great new set of packages,

great updates to existing packages, a lot of those visualization focused, API focused, along with some cool shiny stuff too, if you want to check that out, as well as a recent R Pharma Hangout recording that I was able to have the pleasure of hosting with Daniel Schoberg to talk about the latest and greatest in the world of analysis result datasets and his highly popular GT summary package and the way they work hand in hand together in life science workflows and beyond.

So definitely check that out along with, like I said, the rest of the issue. So, yeah, flight solo here. So far, the flight's still in flight here, so to speak, for the few minutes we have left here. So as always, I'm gonna close out by telling you where you can help the project. That is giving us a link to that great new resource you found online, or maybe you have authored that great blog post, that new package, that great call to action, whatever it is. If it's a way to R and data science, we want to hear about it. We are just a poll request away

on our GitHub repository, which you can see in the upper right corner, the fancy little Octocat. Click on that and you'll be taken straight to GitHub with the template pre populated for you to submit that poll request. We also love to hear from you in terms of what we're doing on this show and just, you know, your adventures in data science. You can use the contact page, which is in this episode's show notes if you want to get in touch with us there, as well as getting in touch with us on social media. You can find me on the, I guess, the big three, if you will. I am on Blue Sky where I am at our podcast at b s k y dot social.

I am also on Mastodon where I'm at our podcast at podcastindex.social, and I'm on LinkedIn, search my name and you will find me there, causing all sorts of stuff. But yeah, conference season is definitely getting underway and I will be at POSICOM, so if any of you have the pleasure of or the good opportunity to go there, please come say hi. I love connecting with the audience. I will be talking about our weekly and our lightning talk that will be hopefully very inspirational

about our past, present, and future of the project. We would love to connect on that, especially as we keep this project going, hopefully, a very long time. Well, this episode has been long enough. Let's close-up shop here. Again, I thank you so much for joining me. Sorry Mike couldn't be here, but hopefully you got some good information out of this anyway, and we hope to have you back for another episode of ROE highlights, hopefully not solo this next time next week.

Open the transcript file

Hosts and Guests

Hosts

  • Eric Nantz 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