Is there a way in geom_sankey to specify an aesthetic that provides directly the size of the flow, i.e. the number of connections between the nodes?

For example:

df <- data.frame(expand.grid(LETTERS[1:3],LETTERS[1:3]))
df$N <- sample(1:10,size = nrow(df),replace = T)

I would like something like

df %>%
make_long(Var1, Var2)%>%
  ggplot( aes(x = x, 
                 next_x = next_x, 
                 node = node, 
                 next_node = next_node,
                 fill = factor(node))) +
  geom_sankey()

But with the flows given by N. A hack would be to repeat each row by N before the make_long, but I am sure there is a proper way.

Following the sum_ issue reported in #1, I have the same issue with the example provided in the readme.

I tried:

library(ggsankey)
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library(ggplot2)
library(gapminder)

df <- gapminder %>%
    group_by(continent, year) %>%
    summarise(gdp = (sum_(pop * gdpPercap)/1e9) %>% round(0), .groups = "keep") %>%
    ungroup()
#> Error: Problem with `summarise()` input `gdp`.
#> x could not find function "sum_"
#> ℹ Input `gdp` is `(sum_(pop * gdpPercap)/1e+09) %>% round(0)`.
#> ℹ The error occurred in group 1: continent = "Africa", year = 1952.

ggplot(df, aes(x = year,
               node = continent,
               fill = continent,
               value = gdp)) +
    geom_sankey_bump(space = 0, type = "alluvial", color = "transparent", smooth = 6) +
    scale_fill_viridis_d(option = "A", alpha = .8) +
    theme_sankey_bump(base_size = 16) +
    labs(x = NULL,
         y = "GDP ($ bn)",
         fill = NULL,
         color = NULL) +
    theme(legend.position = "bottom") +
    labs(title = "GDP development per continent")
#> Error:   You're passing a function as global data.
#>   Have you misspelled the `data` argument in `ggplot()`

^{Created on 2021-04-03 by the reprex package (v1.0.0)}

I also tried removing sum_ and replacing with sum when writing to the variable df, but I also had no luck.

See here:

df <- gapminder %>%
  group_by(continent, year) %>%
  summarise(gdp = (sum(pop * gdpPercap)/1e9) %>% round(0), .groups = "keep") %>%
  ungroup()

ggplot(df, aes(x = year,
               node = continent,
               fill = continent,
               value = gdp)) +
  geom_sankey_bump(space = 0, type = "alluvial", color = "transparent", smooth = 6) +
  scale_fill_viridis_d(option = "A", alpha = .8) +
  theme_sankey_bump(base_size = 16) +
  labs(x = NULL,
       y = "GDP ($ bn)",
       fill = NULL,
       color = NULL) +
  theme(legend.position = "bottom") +
  labs(title = "GDP development per continent")

My console error is different than reprex's for some reason; this is my console error:

Error: Problem with `summarise()` input `flow_freq`.
x could not find function "sum_"
ℹ Input `flow_freq` is `sum_(value)`.
ℹ The error occurred in group 1: n_x = 1952, node = "Oceania 1952", n_next_x = 1957, next_node = "Oceania 1957".

It seems I cannot use ggsankey with my Shiny app.

I did a test app to see if it worked, but I get an error from toJSON about a named vector. I made a simple selector to use a database extract as the base dataframe for the sankey plot. On initialisation, it works, but after switching to another df, I get the error.

Sadly, as an R beginner, I cannot ascertain that I am not at the origin of the issue.

Hi- I just discovered the existence of Sankey plots (or rather, that such things had a name and could be done in R...).

I found your package and ggalluvial, which seems to pre-date ggsankey. Can you comment on the pros and cons of ggsankey? Both packages seem pretty good at a first glance. Thanks!

Thanks for this package! It's really cool. But I saw that under license, you do not have any license. Legally this means no one can use or modify it. Can you add a license?

For more on this you can read this page.

I get this error when I run the example for geom_sankey_bump()

Problem with `summarise()` input `gdp`.
x could not find function "sum_"
ℹ Input `gdp` is `(sum_(pop * gdpPercap)/1e+09) %>% round(0)`.
ℹ The error occurred in group 1: continent = "Africa", year = 1952.
Backtrace:
  1. base::source("~/.active-rstudio-document", echo = TRUE)
 13. base::.handleSimpleError(...)
 14. dplyr:::h(simpleError(msg, call))
Run `rlang::last_trace()` to see the full context.

geom_sankey() works like a charm, by the way! 😀

I have a sankey chart with 21 nodes, and I'm trying to fill the flows one of three colors, but flow.fill isn't working. Is there more documentation on how it works?

Labels always get misaligned. They don't seem to follow any logical rule. Using ggsankey on a W10 laptop with R version 4.2.1. Output was generated with pdf() since there is no proper antialiasing in the png() output (not an issue of ggsankey but of the OS), and some labels get cropped for being so far from the diagram. The package is excellent however!

Suppose I have this dataset (the actual dataset has 30+ columns and thousands of ids)

	df <- data. Frame(id = 1:5,
				admission = c("Severe", "Mild", "Mild", "Moderate", "Severe"),
				d1 = c(NA, "Moderate", "Mild", "Moderate", "Severe"),
				d2 = c(NA, "Moderate", "Mild", "Mild", "Moderate"),
				d3 = c(NA, "Severe", "Mild", "Mild", "Severe"),
				d4 = c(NA, NA, "Mild", "Mild", NA),
				outcome = c("Dead", "Dead", "Alive", "Alive", "Dead"))

I want to make a Sankey diagram that illustrates the daily severity of the patients over time. However, when the observation reaches NA (means that an outcome has been reached), I want the node to directly link to the outcome.

This is how the diagram should look like: []1

Image fetched from the question asked by @qdread here

Is this possible with ggsankey?

This is my current code:

df.sankey <- df %>%
	make_long(admission, d1, d2, d3, d4, outcome)
ggplot(df.sankey, aes(x = x,
					 next_x = next_x,
					 node = node,
					 next_node = next_node,
					 fill = factor(node),
					 label = node)) +
	geom_sankey(flow.alpha = 0.5,
				node.color = NA,
				show.legend = TRUE) +
	geom_sankey_text(size = 3, color = "black", fill = NA, hjust = 0, position = position_nudge(x = 0.1))

Which results in this diagram: []3

Thanks in advance for the help.

in sankey.R, in the function StatSankeyFlow (line 228) is summarise(flow_freq = dplyr::n(), .groups = "keep") which is missing the explicit reference to dplyr.

Hi:

I put it together myself. The coordinates don't match:

This is what I'm looking for:

my code:

library(ggplot2)
library(ggsankey)
library(dplyr)
pl <- ggplot(dat3, aes(x = x, 
                       next_x = next_x,
                       node = node, 
                       next_node = next_node,
                       fill = factor(node),
                       label = node2
                       )) +
  geom_sankey(flow.alpha = 0.5, node.color = "black") +
  geom_sankey_label(size = 6, color = "black", fill = "white", hjust = 1, family = "Times") +
  scale_fill_viridis_d(option = "magma") +
  theme_sankey(base_size = 16) +
  scale_x_discrete(expand = c(0.01,0.1)) +
  theme(legend.position = "none",
        axis.title = element_blank(),
        axis.text = element_blank())
pl

library(clusterProfiler)
kk_dot <- dotplot(kk, showCategory=10) +
  theme(text = element_text(family = "Times"),
        axis.text.y = element_text(size = 12, face = "bold"),
        axis.text.x = element_text(size = 10, face = "bold"),
        axis.title.x = element_text(size = 14, face = "bold"),
        legend.title = element_text(face = "bold"))
kk_dot
kk_dot2 <- kk_dot + theme(axis.text.y = element_blank(),
                          axis.ticks.y = element_blank())
library(patchwork)
design <- c("
            AAAA#
            AAAAB
            AAAAB
            AAAAB
            AAAA#
            ")
all_p <- pl + kk_dot2 + theme(text = element_text(size = 20), 
                              axis.title.x = element_text(size = 25),
                              axis.text.x = element_text(size = 20)) +
  plot_layout(design = design)
all_p

Looking forward to your reply!