demo/bar.r

In rpruim/ggvis2: Interactive grammar of graphics

library(ggvis)

# Bar graph with continuous x, and y value supplied in the data
pressure %>% ggvis(x = ~temperature, y = ~pressure) %>%
  layer_bars()

# Categorical x, and y var supplied
pressure %>% ggvis(~factor(temperature), ~pressure) %>% layer_bars()

# No y var, and continuous x: bar graph of counts
mtcars %>% ggvis(x = ~cyl) %>% layer_bars()

# Notice how it differs from a histogram: a histogram has bins that span
# ranges of x, but layer_bars shows the count at each unique x value.
mtcars %>% ggvis(~wt) %>% layer_histograms()
mtcars %>% ggvis(~wt) %>% layer_bars()

# No y var, and categorical x: bar graph of counts at each x value
mtcars %>% ggvis(~factor(cyl)) %>% layer_bars()


# Hair and eye color data
hec <- as.data.frame(xtabs(Freq ~ Hair + Eye, HairEyeColor))

# Without stacking - bars overlap
hec %>% group_by(Eye) %>%
  ggvis(x = ~Hair, y = ~Freq, fill = ~Eye, fillOpacity := 0.5) %>%
  layer_bars(stack = FALSE) %>%
  scale_nominal("fill",
    domain = c("Brown", "Blue", "Hazel", "Green"),
    range = c("#995522", "#88CCFF", "#999933", "#00CC00"))

# With stacking
hec %>% group_by(Eye) %>%
  ggvis(x = ~Hair, y = ~Freq, fill = ~Eye, fillOpacity := 0.5) %>%
  layer_bars() %>%
  scale_nominal("fill",
    domain = c("Brown", "Blue", "Hazel", "Green"),
    range = c("#995522", "#88CCFF", "#999933", "#00CC00"))


# Stacking in x direction instead of default y - need to be explicit about
# all the steps
hec %>% group_by(Eye) %>%
  ggvis(y = ~Hair, fill = ~Eye, fillOpacity := 0.5) %>%
  compute_stack(stack_var = ~Freq, group_var = ~Hair) %>%
  layer_rects(x = ~stack_lwr_, x2 = ~stack_upr_, height = band()) %>%
  scale_nominal("y", range = "height", padding = 0, points = FALSE) %>%
  scale_nominal("fill",
    domain = c("Brown", "Blue", "Hazel", "Green"),
    range = c("#995522", "#88CCFF", "#999933", "#00CC00"))