Matthew Kay

#ggdist 3.3.0 just hit CRAN! Some highlights:

1. A big change is the new default density estimator, which detects bounded distributions; see linked post: https://fediscience.org/@mjskay/110359116199194043

This also means Mode() and hdi(), which use the density estimator, should be better too.

#rstats

Matthew Kay (@[email protected])

Attached: 4 images round 2 of "try to make #ggdist's default density estimator better" #rstats I think this is the one for now (better be too b/c gotta release another version soon!!!)

FediScience.org
May 14, 2023 at 1:23amWeb
Show threadMatthew KayMay 14, 2023
2. Histogram support is also improved through a customizable histogram density estimator, which better supports fixed bin width and aligning bins to specific points (inspired by similar changes in ggplot2::geom_histogram, but with a different take on the API).
Show threadMatthew KayMay 14, 2023
3. The new geom_spike() / stat_spike() make it easy to annotate specific values on a slab; see https://mjskay.github.io/ggdist/articles/slabinterval.html#annotating-slabs-with-spikes
Slab + interval stats and geoms

ggdist

Show threadMatthew KayMay 14, 2023
4. Finally, ggdist::parse_dist() now respects the lower and upper bounds of priors from @paul_buerkner 's brms package, which makes it easier to visualize things like truncated Normal priors (sigma in the figure); see https://mjskay.github.io/ggdist/articles/slabinterval.html#visualizing-priors
Slab + interval stats and geoms

ggdist