ggalluvial ordering stratum - r

Question

I've been trying to make some sankey diagrams with the ggalluvial package. I rather like it but I'm having problems controlling the order of the lodes. I'm using the alluvia format described at the start of the vignette.
Basically my diagram is showing subsets of level 2 and level 3 of one time point and how they move to another time point. The problem is that I can't for the life of me figure out how to force the order of the stratum as the diagram is unreadable without the order being correct. Here's my code:
library("ggalluvial")
library("ggplot2")
subsank_math = structure(list(`Winter Projection` = structure(c(2L, 2L, 2L,
2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L, 4L, 4L, 5L,
5L, 5L, 5L, 5L, 5L), .Label = c("Level 5", "Level 4", "Level 3",
"Level 2", "Level 1"), class = "factor"), subgroup = structure(c(1L,
2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L, 6L, 1L, 2L, 3L, 4L, 5L,
6L, 1L, 2L, 3L, 4L, 5L, 6L), .Label = c("Level 3 (+)", "Level 3",
"Level 3 (-)", "Level 2 (+)", "Level 2", "Level 2 (-)"), class = "factor"),
n = c(119, 102, 16, 10, 12, 1, 272, 544, 182, 151, 134, 22,
40, 239, 204, 326, 663, 225, 0, 15, 12, 44, 215, 219)), row.names = c(NA,
-24L), .Names = c("Winter Projection", "subgroup", "n"), class = "data.frame")
ggplot(subsank_math,
aes(weight = n,
axis1 = subgroup, axis2 = `Winter Projection`)) +
geom_alluvium(aes(fill = subgroup),
width = 0, knot.pos = 0, reverse = FALSE) +
geom_stratum(width = 1/8, reverse = FALSE) +
geom_text(stat = "stratum", label.strata = TRUE, reverse = FALSE)
Kinda strange that it orders the first axis according to the levels of the factor but not the second.

Answer 1

I've just started playing with the ggalluvial package myself, so I won't claim to understand how things work, but reformating the data frame into lode format (described near the end of the package's vignette) worked for me:
library(dplyr)
library(tidyr)
df.lode <- subsank_math %>%
mutate(subject = seq(1, n())) %>%
gather(x, level, -n, -subject) %>%
mutate(level = factor(level,
levels = c("Level 1", "Level 2 (-)", "Level 2",
"Level 2 (+)", "Level 3 (-)", "Level 3",
"Level 3 (+)", "Level 4")))
> head(df.lode)
n subject x level
1 119 1 Winter Projection Level 4
2 102 2 Winter Projection Level 4
3 16 3 Winter Projection Level 4
4 10 4 Winter Projection Level 4
5 12 5 Winter Projection Level 4
6 1 6 Winter Projection Level 4
ggplot(df.lode,
aes(x = x,
stratum = level,
alluvium = subject,
weight = n,
label = level)) +
geom_flow(aes(fill = level)) +
geom_stratum() +
geom_text(stat = "stratum") +
scale_fill_discrete(limits = levels(df$a1))