gganimate with multiple points for same x value

Question

I have some data in which multiple y coordinate values for a single x coordinate value. It is similar to the one included in this question. It also has a state variable which has to be used for the transition_state() function in gganimate.

library(data.table)

df <- data.table(
  x = seq(11),
  type1 = c(2, 3, 2.5, 3, 2, 2, 3, 2.5, 3.5, 3, 2),
  type2 = c(NA, 3, 3.5, 3, NA, NA, 3, 3.5, 4, 3, NA)
)
m.df <- melt.data.table(df, "x", variable.name = "grp")
m.df[x %in% seq(5), "state" := 1][x %in% seq(6, 11), "state" := 2]
m.df
#>      x   grp value state
#>  1:  1 type1   2.0     1
#>  2:  2 type1   3.0     1
#>  3:  3 type1   2.5     1
#>  4:  4 type1   3.0     1
#>  5:  5 type1   2.0     1
#>  6:  6 type1   2.0     2
#>  7:  7 type1   3.0     2
#>  8:  8 type1   2.5     2
#>  9:  9 type1   3.5     2
#> 10: 10 type1   3.0     2
#> 11: 11 type1   2.0     2
#> 12:  1 type2    NA     1
#> 13:  2 type2   3.0     1
#> 14:  3 type2   3.5     1
#> 15:  4 type2   3.0     1
#> 16:  5 type2    NA     1
#> 17:  6 type2    NA     2
#> 18:  7 type2   3.0     2
#> 19:  8 type2   3.5     2
#> 20:  9 type2   4.0     2
#> 21: 10 type2   3.0     2
#> 22: 11 type2    NA     2
#>      x   grp value state

The data is plotted using gganimate as shown below.

library(ggplot2)

ggplot(m.df, aes(x, value, group = grp, color = grp)) +
  geom_line(na.rm = T) +
  geom_point(na.rm = T) +
  theme_bw()

The plot can be found HERE

I want to use the gganimate transition_state() function using the state column in the data as shown below, but shows an error.

library(gganimate)

ggplot(m.df, aes(x, value, group = grp, color = grp)) +
  geom_line(na.rm = T) +
  geom_point(na.rm = T) +
  transition_states(state)
#> Error in rep(seq_len(nrow(polygon)), splits + 1): invalid 'times' argument

What am I doing wrong here?

Thank you in advance

Answer 1

One option to make your code work is to group by both grp and state using e.g. interaction. Otherwise gganimate fails to split your data by state as both states are present in each group. Additionally, you have to put the group aes after the color aes. Otherwise you run in to the same issue:

library(ggplot2)
library(gganimate)

ggplot(m.df, aes(x, value, color = grp, group = interaction(grp, state))) +
  geom_line(na.rm = T) +
  geom_point(na.rm = T) +
  transition_states(state)