LoginSign Up
← Back to all sparks
G

ggpubr

ANALYTICS
Velocity0.0

Publication-ready plots in R with simplified ggplot2 syntax and statistical annotations

github.com/kassambara/ggpubr ↗

ggpubr reached 1.0.0 with p-value formatting presets for specific journals

visualizationstatisticspublicationggplot2deprecation
◆Current state
ggpubr adds publication-ready statistics and annotation to ggplot2. Two releases define its capability: 0.5.0 introduced the stat_*_test family and geom_pwc() for pairwise comparison brackets, and 1.0.0 added p-value formatting presets matching named journal house styles. In between, most releases are ggplot2 and dplyr deprecation chasing.
◆Where it's heading
The package is moving from drawing statistics to matching the conventions of where they get published - style presets are a different kind of feature from a new test. That sits on a persistent maintenance load: after_stat migrations, linewidth parameters, R-devel changing how the Wilcoxon test handles ties. ggpubr absorbs upstream deprecations so that figure code written years ago keeps rendering.
◆Prediction
Expect the preset list to grow as users request their own journals' conventions, and the deprecation-chasing to continue with each ggplot2 release; the statistical test coverage looks complete enough that additions there would be surprising.

◆Recent moves

  1. 1mo ago

    Journal-specific p-value formatting presets in 1.0.0

    ⚡ SPARK

    The 1.0.0 milestone shifts what the package optimizes for: not what statistic to draw, but how a target journal expects it written. It builds directly on the annotation machinery 0.5.0 introduced, adding placement and orientation controls to the same brackets and labels.

    View source ↗
  2. 5mo ago

    Raises R and dplyr minimums, migrates off deprecated syntax

    A cleanup release that pays down years of accumulated deprecation: dot-dot notation replaced with after_stat(), dplyr::do() with reframe(), tidyr::gather() with pivot_longer(), and size superseded by linewidth. The version floors move to match what ggplot2 already required.

    View source ↗
  3. 9mo ago

    Pins Wilcoxon p-values against an R-devel change

    R-devel changed the Wilcoxon test to give exact conditional inference with ties, which would have silently shifted reported p-values, so compare_means() now sets exact = FALSE to keep results consistent across R versions. A compatibility fix, but one that protects published numbers rather than just checks.

    View source ↗
  4. 1y ago

    Fixes after_stat() namespace failures in reverse dependencies

    The automatic conversion from dot-dot notation to after_stat() was resolving namespaces incorrectly and breaking packages that build on ggpubr. Also adds an outliers argument to ggboxplot() and corrects the regression equation to the conventional y = mx + b ordering.

    View source ↗
  5. 3y ago

    ggadjust_pvalue() and reproducible jitter seeds

    Adds multiple-comparison adjustment for p-values already drawn by geom_pwc(), completing the annotation workflow from 0.5.0. The seed handling changes matter more than they read: jittered points and repelled labels now restore the random state instead of consuming it.

    View source ↗
  6. 3y ago

    geom_pwc() and the stat_*_test family arrive

    ⚡ SPARK

    The release that established how ggpubr annotates significance, and the base every later release extends. Everything since - p-value adjustment in 0.6.0, anchoring and orientation in 1.0.0 - refines this machinery rather than replacing it.

    View source ↗