home.social
  1. RE: fosstodon.org/@terence/1165730

    I can’t emphasize how much I like the legend that also tells you the proportion of the whole the variable represents! It’s quite difficult to parse proportions with such a disconnected set, so this is a clever way of including it in a figure.

  2. ggplot(ritual) +
    geom_sf() +

    theme_c̵̤͛̌̎̀͘t̶̡̨̯̝̼͗̂͒͊̀̾̇̾͜h̸̡̰͉̼̖͉̉u̶̡̗̮̜͍̠̯̙̍l̸̞̲̓̒̒̓̓̃̔h̶̛̖̣̲̩̖͉̠͍̬̄̔̊̽͜u̸̱͛̈̽̀̅() +

    coord_noneuclidean()

  3. why would I get AI to write my documentation and examples, that's the best part.

  4. Get in, kids! An ADSB storm is coming!

  5. Want dynamically-generated clickable links in the Positron console to your {targets} pipeline, across all your files, so you can immediately jump to an error or jump to a long-running target and inspect it? 🌳tarborist has your back!*

    *as long as you're okay building a custom version of Positron from source with a completely new ConsoleLinkProvider API😜

    github.com/tylermorganwall/pos

  6. I released an update for 🌳tarborist last night! Along with bugfixes, this version brings more info on the depth of child nodes in the hover pane and DAG distance in the file picker. Most importantly, you can now hover over a target in your {targets} pipeline and get a link to jump right to it—even if it's in an entirely different file. Make navigating the pipeline much quicker.

  7. Update: It's not a true post until something is spinning in 3D 😀

  8. Just like Starbucks and Dunkin', even the rayverse is getting on the Protein-everything bandwagon :)

  9. 3D ggplot of parabolic great circle trajectories (i.e. ballistic-ish paths) on a non-linear Robinson projection: how do we ensure the 3D data is consistent with the underlying ggplot? Simple: we extract the coordinate transformation from the ggplot object itself!

  10. Well well well, what do we have here? Is that... real 3D data in my 3D ggplot? 🤫

  11. An important question from my latest blog post:

    "It would be nice to have a clean, customizable environment map that gives us that pristine golden-hour light without any real-world interlopers. Is this a realistic goal–or is it just a pie in the sky dream?"

    Post: tylermw.com/posts/rayverse/atm

  12. And here's the hex logo! And unlike Bluesky, Mastodon actually respects PNG transparency!

  13. 3/5 We can use this information along with solar positioning via {swephR} to generate a beautifully rendered realistic atmosphere in our and plots! Rayshader's `render_highquality()` now optionally takes lat/long + datetime inputs, so you don't even need to learn anything new!

  14. 2/5 implements the Prague (cgg.mff.cuni.cz/publications/s) and Hosek (cgg.mff.cuni.cz/projects/Skyli) atmospheric models, which allow you to generate skies with the sun + scattered atmospheric light for any solar angle, down to -4.2 degrees (sun below the horizon) for the Prague model!

  15. 1/5‼️Gigantic new update + blog post! Introducing {skymodelr}: render 3D scenes in with realistically-lit skies at a given location, using just a latitude, longitude, and a time.

    Blog:
    tylermw.com/posts/rayverse/atm

    Github:
    github.com/tylermorganwall/sky

    Site:
    www.skymodelr.com

  16. Happy Groundhog Day! In honor of the day, here's an infinitely repeating video of a realistic atmospheric simulation showing the shadows the sun casts on February 2nd. (over Monterey Bay, California, because no one would want to go to Punxsutawney, PA right now).

  17. An "atmospheric" frame from an animation demonstrating the "Mahattanhenge" phenomenon in my upcoming blog post about rendering realistic atmospheres in R.

    What's fun about rendering a long animation overnight is it makes each morning a miniature Christmas: you never know what you're going to unwrap! Sometimes it's wonderful, and sometimes you forget to increment the counter so you end up rendering the same frame over and over again 😉

  18. Help! I'm trying to write a blog post but I've fallen into the customizing-my-quarto-blog quicksand! The more I struggle to get out the deeper I sink! I've written more CSS for this R blog post than R!

  19. the big dippeR

    (rendered right before dawn at 45,000 ft using a fisheye lens)

  20. apropos of nothing: using R to render a mastodon to post on mastodon

    (this post made more sense on bluesky)

  21. R: the language that sets the gold standard for dataviz!

    Anonymous user K. Midas says: "I love it!"

    (from a cheeky figure in an upcoming blog post about rendering 3D plots with "golden hour" lighting by generating synthetic sunsets, entirely in with )

  22. ‼️ Big rayimage update (v0.23.0)! In addition to 'JPEG', 'PNG', 'TIFF', and 'EXR' files, `ray_read_image()` now supports loading 'TGA', 'BMP', 'PSD', 'GIF', 'HDR', 'PIC', and 'PNM' files (thanks to stb_image): It's a one-stop shop for bringing image data into R.

    GH:
    github.com/tylermorganwall/ray

  23. Do you know it's been almost 7 years since I released , ranked for 6 years straight as the #1* pathtracer in ? I went back to a blog post (tylermw.com/posts/rayverse/get) I wrote over six years and recreated some of the scenes, and it's fun seeing how far it's come! The original images took a long time to render and suffered from severe noise, but years of sampling improvements, BVH traversal efficiencies, and denoising support has made decent renders almost instantaneous!

    *out of 1

  24. You can see the difference in the JFA by looking at the areas where the bonds and atoms intercept here in a molecule rendered with . The old method worked by placing an inverted mesh over the mesh, which results in issues where other meshes intersect (note the black lines where the spherical edges of the atoms intersect with the bonds--those are unwanted visual artifacts from the inverse mesh trick that are gone with the new method).

  25. ‼️ Rayverse package update: rayvertex v0.14.0 released! This overhauls the toon shading algorithm to use a screen space jump fill algorithm (JFA), removing the brittle inverse mesh hack that existed up until now.

    github.com/tylermorganwall/ray

  26. Manhattanhenge: On May 28th/July 13th, the sun aligns with Manhattan's street grid 29° from N. This twice-a-year alignment of urban planning and orbital motion serves both as a temporal tourist attraction as well as a particular good test case for geospatial skymap alignment!

  27. Rendering the peak of Mount Everest from sunrise to sunset in R with rayshader and rayrender

  28. 1) Keeping the exposure the same as the dawn image at noon: ooph!

    2) Adjusting the exposure down three stops with rayimage::render_exposure(): Nice!

    Also: this is why you use a high dynamic range format like EXR when rendering, which rayrender now supports! And also why I spent six months getting all the packages to install OpenEXR on the CRAN, so this magic can occur.

  29. Rendering a tilt-shifted Mount Everest at dawn:

    (dawn and dusk are the "magic hour" in rendering too, because you can represent an image in the dynamic range of a PNG/JPEG without clipping or tonemapping hacks!)

  30. A set of new render_highquality() options coming to a near you 😀

  31. Day 6 of the (3D), a render of Mount Rainier with a realistic atmospheric skydome generated w/ a lat/long/datetime, rendered entirely in with and ! The atmosphere is annoyingly realistic: I now have to dial in the exposure depending on the time of day! Thank goodness for 's render_exposure() function :)

    (someone really needs to create an HDR R graphics device...)

  32. Big cool features coming to the : it only took a half year of CRAN ping pong for 3 low level infra packages with challenging configurations and a complete internal overhaul of rayimage to support proper colorspace management, all to just set the stage for the actual cool stuff :)

  33. Happy Halloween!

    @terence inspired me to whip up an animation: a pumpkin made entirely in rayrender! (CSG for the carved pumpkin, extruded path stem, cyl + sphere candle, and r_obj eyes)

    Also, denoising made the below animation render in <3 minutes!

    Code:
    gist.github.com/tylermorganwal

  34. Pro-tip: Never learn about colorspaces. "I love learning new things. Why yes, I'll eat that red apple, what a friendly snake!"

    ::crunch::

    "Wait... this apple isn't , it's outside the sRGB gamut! Why does everything look terrible? Oh no... do I have to fix years of ad-hoc color handling?"😱

    ...

    "Anyway, God then kicked me out of the 8-bit sRGB Garden of Eden after I kept pointing out banding and muddy color transitions in his rainbows"

  35. 5/5 Focusing back on a single lat/long, here we simulate the analemma across the day in DC (here, in 30 minute increments) to see how the shape changes position across the sky/year. Basically, visualizing a 2D projected slice of 4D time/space: that's a fun !

  36. 1/5 Modeling the atmosphere at noon in Washington, DC for a full year in with the and my upcoming sky modeling package. If you follow the path of the sun, you might notice that it travels in a lopsided figure eight curve. Why? What's going on here?

  37. Woo hoo! The nloptr package is now supported in webR, which means we now have access to the lme4 package for mixed effects modeling! This brings a great deal of statistical modeling power to serverless R workflows. Great work @gws !

  38. ‼️New update: v0.17.1 has been released, and with it comes some cool improvements:

    * R console ANSI full-color preview if supported (Positron does on my system!)
    * Pretty printing for RGBA arrays!
    * Proper gamma correction tracking and EXR support

    Site:
    www.rayimage.dev

  39. Testing the accuracy of simulating star trails with R and libopenexr, rendered in a pathtracer: lookin' pretty good so far! Differences seem to be primarily from real camera lens distortion from the wide angle lens versus a perspective camera.

  40. Procedurally rendered location/time-accurate star trails + simulated twilight atmosphere, now with star spectral → RGB color information included! It's subtle (even exaggerated here slightly), but there. All rendered in with

  41. 🤯 Rendering realistic sunsets in R
    🤯 🤯 Rendering realistic sunsets in R, with accurate stars for a given latitude/long
    🤯 🤯 🤯 Rendering realistic sunsets in R, with accurate stars for a given lat/long AND time/exposure, giving you beautiful star trails!

    your move, SAS

  42. Nothing is better than relaxing on a hot summer Friday night, sipping a cold drink, and enjoying a beautiful sunset accurately rendered (as if seen from a jet airliners cruising altitude while on a red eye) in your favorite statistical programming language 🥂

  43. Twilight support!

    One of my favorite things is finding a methods paper (here, atmospheric rendering), implementing it, thinking to myself "Man, it would be nice if the method supported XYZ!", and then finding that the paper implementing XYZ was released fairly recently. Truly surfing the crest of human knowledge!

  44. After setting up the required infrastructure on CRAN with several packages providing static libraries, I finally have some new features on the way! Shown here: dynamic sky generation for a given lat/long and time of day, based on the Hosek-Wilkie radiance model.

  45. Great 3D of data (!) on DIB on the distribution of frags in the classic Counterstrike map, de_dust2. The author followed a tutorial by @mrpecners.bsky.social (not on mastodon) and commented that the hard part was getting the data--the 3D rendering was trivial!👍 Shows the power of a good tutorial and !😀

    reddit.com/r/dataisbeautiful/c

  46. Who says writing documentation isn't fun?😀🐝

  47. 1/3 ‼️Package update alert‼️Need a doomscrolling break? How about a dramatically faster version of to clear your mind! Developed using the nice C/C++ features of the new @Posit Positron IDE, rayrender is more than 40% faster than before! Less heat, less time, faster 3D data viz!