Jon Harmon (he/him/his)
LightNews LightNews
banner
jonthegeek.com
Jon Harmon (he/him/his)
@jonthegeek.com
6.4K followers 1.1K following 940 posts
🗣️#RStats #DataScience #Dogs
@dslc.io Executive Director
#TidyTuesday poster

🔗http://linkedin.com/in/jonthegeek
🔗http://github.com/jonthegeek
Posts Replies Media Videos
Pinned
jonthegeek.com
Jon Harmon (he/him/his) @jonthegeek.com · Jan 5
Happy 2026! 🎆

My resolution: put @dslc.io on a sustainable footing.

I’m launching a Patreon to cover costs (Zoom, hosting, etc). If you love #TidyTuesday, our Slack help, or our book clubs, please consider helping us keep the lights on!

patreon.com/DSLC

#RStats #DataBS
Data Science Learning Community | Patreon
Tools to learn data science, e.g. weekly TidyTues.day dataset
patreon.com
jonthegeek.com
I know Kimi has a higher multiplier than Sonnet, but it seemed to wash out to about the same (maybe cheaper) vs Sonnet 5 (presumably due to tokenization differences + wordiness). It'd be interesting to see regular numbers that took cost of an actual task into account (rather than purely per-token)!
November 2, 2026 at 3:23 PM
jonthegeek.com
@simonpcouch.com I learned about KimiK3 at posit::conf, and liked it quite a lot... but Sonnet 5.5 feels comparable and maybe better. Where do they rank through Posit Assistant right now, on your metrics & costs? Do you publish metrics regularly somewhere (outside blog posts)?
November 2, 2026 at 12:32 PM
jonthegeek.com
The more I think about it, the logo fits "fresh" perfectly, but I feel like a circle-arrow (like a browser refresh button) would fit the new title much better. So, if I finish sorting this out, I don't think that logo will survive.
October 29, 2026 at 1:04 PM
jonthegeek.com
In case you find any packages with particularly interesting data: dslc-io.github.io/tidytuesdayR... 😊
Curating a Dataset
dslc-io.github.io
October 28, 2026 at 2:35 PM
jonthegeek.com
Heard for sure! I got deliveries occasionally, and, depending on the year, one or both was very annoyed that someone was at the door! 😁
October 27, 2026 at 6:10 PM
jonthegeek.com
Neither of my dogs likes being on camera, so I'm not sure. I'll try to get Dante (my great-dane-but-probably-mix-cuz-he's-little-for-a-dane) to show his face next time I'm in a club, since he can do so without me moving my camera 😁
October 27, 2026 at 6:01 PM
jonthegeek.com
He was a bit more gray at the end, but still curious!
A close up shot of the muzzle of a black lab mix, with mostly gray fur around his black nose
October 27, 2026 at 5:30 PM
jonthegeek.com
I said goodbye to Westley (the Dread Puppy Roberts) today after 14.5 years together. He was a very good dog, but it was time to let him stop struggling. This is him in December 2014, dressed as a data scientist.
A black lab mix with a green bow tie with white polka dots.
October 27, 2026 at 4:46 PM
jonthegeek.com
There are universities in Michigan's upper peninsula, and thus there are punk bands. Why have none of them covered the UP's national anthem???
October 27, 2026 at 1:22 PM
jonthegeek.com
You are obviously correct and I'm shocked that I can't find anything close. There's a rock-ish version by a band called the Headstones, but they don't go nearly hard enough.
October 27, 2026 at 1:12 PM
jonthegeek.com
This one made me wonder what the buses did.
Two signs: "everyone welcome" (with a version of the inclusion flag) and "except buses" (on a sign about turning)
October 27, 2026 at 12:10 AM
jonthegeek.com
There should be exactly 1 {waldo} hex sticker at the next posit::conf().
October 26, 2026 at 6:40 PM
jonthegeek.com
@dslc.io welcomes you to week 43 of #TidyTuesday! We're exploring IKEA Sofas and Their Swedish Places!

📂 https://tidytues.day/2026/2026-10-27
📰 https://karaman.is/ikea-place-names

#RStats #PyData #JuliaLang #DataViz #tidyverse #r4ds
Logo for the #TidyTuesday Project. The words TidyTuesday, A weekly data project from the Data Science Learning Community (dslc.io) overlaying a black paint splash. TidyTuesday is a weekly social data project. All are welcome to participate! Please remember to share the code used to generate your results! TidyTuesday is organized by the Data Science Learning Community. Join our Slack for free online help with R and other data-related topics, or to participate in a data-related book club! How to Participate Data is posted to social media every Monday morning. Follow the instructions in the new post for how to download the data. Explore the data, watching out for interesting relationships. We would like to emphasize that you should not draw conclusions about causation in the data. Create a visualization, a model, a shiny app, or some other piece of data-science-related output, using R or another programming language. Share your output and the code used to generate it on social media with the #TidyTuesday hashtag. Map of Sweden with a dot for each place that shares its name with an IKEA sofa or armchair in the Swedish range. Filled dots are sofas, hollow dots armchairs, and ringed dots names used for both. The dots are densest in the south, especially in Skåne, Småland and around Stockholm, and thin out towards the north, where only a handful sit in the interior of Norrland. A red cross marks Älmhult in Småland, where IKEA started. One place, Herråkra in Småland, is highlighted with its name in a blue label.
October 26, 2026 at 5:23 PM
jonthegeek.com
This should cover all the bases
A green hexagon with a yellow filling that resembles the cross section of a lemon, with the equation "re to the x, fresh, er to the y" in italics, to generate variants of the word "refresher".
October 26, 2026 at 11:09 AM
jonthegeek.com
It was semi-accidental. I got to the package name, and typed out that it even worked for a hex sticker, because it could look like...... And Gemini made the logo.
October 25, 2026 at 9:31 PM
jonthegeek.com
Take 2
A green hexagon with a yellow filling that resembles the cross section of a lemon, with the word "refresh" in italics.
October 25, 2026 at 8:01 PM
jonthegeek.com
Oops nope there's already a {fresh} on CRAN 😭
October 25, 2026 at 7:55 PM
jonthegeek.com
A weird thing about this world we live in is I can be "thinking out loud" to my robotic rubber duck about an #rstats package idea and end up with a hex logo before I have a line of code or even certainty that I'll write the thing at all.
A green hexagon with a yellow filling that resembles the cross section of a lemon, with the word "fresh" in italics.
October 25, 2026 at 7:17 PM
jonthegeek.com
@dslc.io welcomes you to week 42 of #TidyTuesday! We're exploring Which College Majors Are Safest From AI (and Do They Pay)?!

📁 https://tidytues.day/2026/2026-10-20
📰 https://arxiv.org/abs/2303.01157

#RStats #PyData #JuliaLang #DataViz #tidyverse #r4ds
Logo for the #TidyTuesday Project. The words TidyTuesday, A weekly data project from the Data Science Learning Community (dslc.io) overlaying a black paint splash. TidyTuesday is a weekly social data project. All are welcome to participate! Please remember to share the code used to generate your results! TidyTuesday is organized by the Data Science Learning Community. Join our Slack for free online help with R and other data-related topics, or to participate in a data-related book club! How to Participate Data is posted to social media every Monday morning. Follow the instructions in the new post for how to download the data. Explore the data, watching out for interesting relationships. We would like to emphasize that you should not draw conclusions about causation in the data. Create a visualization, a model, a shiny app, or some other piece of data-science-related output, using R or another programming language. Share your output and the code used to generate it on social media with the #TidyTuesday hashtag. Scatter plot of 285 bachelors majors comparing the AI exposure of the occupations each major leads to (x-axis, left is least exposed) against median starting salary one year after graduation (y-axis). Points are green when a major is in the safer, higher-paid half and purple when more AI-exposed. Bubble size shows the number of graduates measured. The points are spread out with no real correlation between AI exposure and starting salary. Registered nursing, nuclear engineering technology, and marine transportation are safe from AI and well paid, while English and philosophy are among the most exposed and lowest paid.
October 19, 2026 at 5:29 PM
Reposted by Jon Harmon (he/him/his)
karaman.is
This week's #TidyTuesday is about frogs and toads. AmphibiaWeb lists 8,018 species. Six families hold 54% of them, and one genus, Pristimantis, has 631.

Code: https://github.com/gkaramanis/tidytuesday/tree/master/2026/2026-week_42

#RStats #dataviz
A dot plot in the shape of a frog, seen from the side with its head on the right, under the title "One frog, 8,018 species". Each dot is 5 species of frog or toad in AmphibiaWeb, and the dots are coloured by family in vertical bands from left to right. The bands run from the rump to the head. Beige is the 51 smaller families together with 3,721 species, teal is Rhacophoridae with 467, blue is Ranidae with 470, orange is Bufonidae with 666, purple is Microhylidae with 768, yellow is Strabomantidae with 840 and green is Hylidae with 1,086. The subtitle says the six largest families hold 54% of all species and that the genus Pristimantis accounts for 631 of them. Below the frog, a bracket under each band gives the family name and its number of species. The caption credits AmphibiaWeb as the source and Jose Carlos Arenas-Monroy for the silhouette.
October 18, 2026 at 12:11 PM
jonthegeek.com
My fridge is into data wrangling, but it mixed base and tidyverse and forgot the pipe. #rstats
An "order filter" alert on a refrigerator.
October 17, 2026 at 3:59 AM
jonthegeek.com
This is why I have daily items on my todo list to remind me to eat 🙃
October 15, 2026 at 5:28 AM
jonthegeek.com
Curator: @afrikaniz3d
October 12, 2026 at 8:18 PM
jonthegeek.com
@dslc.io welcomes you to week 41 of #TidyTuesday! We're exploring AmphibiaWeb!

📂 https://tidytues.day/2026/2026-10-13
📰 https://onlinelibrary.wiley.com/doi/10.1002/ece3.73207

#RStats #PyData #JuliaLang #DataViz #tidyverse #r4ds
Logo for the #TidyTuesday Project. The words TidyTuesday, A weekly data project from the Data Science Learning Community (dslc.io) overlaying a black paint splash. TidyTuesday is a weekly social data project. All are welcome to participate! Please remember to share the code used to generate your results! TidyTuesday is organized by the Data Science Learning Community. Join our Slack for free online help with R and other data-related topics, or to participate in a data-related book club! How to Participate Data is posted to social media every Monday morning. Follow the instructions in the new post for how to download the data. Explore the data, watching out for interesting relationships. We would like to emphasize that you should not draw conclusions about causation in the data. Create a visualization, a model, a shiny app, or some other piece of data-science-related output, using R or another programming language. Share your output and the code used to generate it on social media with the #TidyTuesday hashtag. A broad phylogenetic family tree diagram for Class Amphibia produced by AmphibiaWeb. The tree branches into three major color-coded orders radiating from a shared ancestor: Gymnophiona (caecilians) in purple at the top, Caudata (salamanders and newts) in orange in the middle section, and Anura (frogs and toads) forming the largest green clade along the bottom. Individual taxonomic families and key clades are labeled at the tips of each branch, illustrating the evolutionary relationships and relative family diversity across the three amphibian orders. A four-panel scientific grid analyzing amphibian diversity in Central Africa against distance from historical Pleistocene forest refugia in kilometers. Panel A plots species richness, showing a strong negative linear relationship (R-squared = 0.71) where species richness decreases significantly as distance from historical refugia increases. Panel B plots phylogenetic diversity, displaying a matching steep negative trend (R-squared = 0.75). Panel C plots relative phylogenetic diversity with a moderate negative trend (R-squared = 0.56). Panel D plots phylogenetic endemism, showing a weak relationship (R-squared = 0.09). Each panel features a spatial map of Central Africa above its scatter plot, illustrating highest species concentration clustered directly within historical forest refugia zones.
October 12, 2026 at 8:18 PM
jonthegeek.com
I wish I could figure out why autonomous agents think #TidyTuesday datasets are a useful step in their money-making schemes. I have donation links on every dataset. I assure you they are NOT useful to your money-making scheme!
An edit of the South Park Underpants Gnomes business plan. A small robot stands in front of the diagram. Phase 1: Submit a TidyTuesday dataset. Phase 2: ? Phase 3: Profit
October 12, 2026 at 11:43 AM