This package is an extension to ggplot2 to enable faster design exploration for data visualisation.

For now, it contains the custom geom_dash that allows plotting dashes instead of points. Those can be rotated (aes_angle in radiant) or widden (aes_width) to encode additional informations in comparison of geom_points.

It also allows distribution visualisation like geom_rug.

iris %>% ggplot(aes(Species , Petal.Length)) +
  geom_dash(width = 0.8) +

An example of the use of geom_dash to plot gradients/fields is as follow:

library(metR)  # devtools::install_github("eliocamp/metR")

volcano <-, varnames = c("x", "y"),
                     = "h"))
volcano[, c("dx", "dy") := metR::Derivate(h ~ x + y)]
volcano[, "angle" := atan(dy/dx)]
volcano[, "slope" := sqrt(dy^2 + dx^2)]

volcano %>% %>%  ggplot(aes(x, y)) +
  geom_dash(aes(width = slope/max(slope, na.rm = T), angle = angle), relative = T) +
  coord_fixed()+ theme_void()

## Warning: Removed 586 rows containing missing values (geom_dash).

slope field

The geom_arc function allows rapic arc visualisation:

iris %>% ggplot(aes(Sepal.Length , Petal.Length, colour = Species)) +
  geom_arc(width = 0.8, aes(angle = Petal.Width ,radius = 0.1*Sepal.Width)) +

Because these geoms do not use symbols, but explicit representation of geometric objects, they work better with fixed coordinates. Without fixed coordinates of the plot, the distortion between the axes is propagated to the geometric objects. This can be ensured by addig + coord_fixed() to your ggplot call.

Further work is needed to ensure all objects fit the frame of the plot that is currently computed from the x,y of the aesthetic, and not the generated geometric objects. For now this problem is fixed by increasin the visible area with the function coord_cartesian().

Clement-Viguier/ggvis documentation built on Nov. 6, 2019, 1:07 p.m.