Plotly Sankey finetuning; node alignment along x-axis, drop-off

Question

The following diagram is close to what I am looking for, however I would like to know if the following is possible:

• Left-alignment of nodes rather than justified along the x-axis?, so for e.g., a flow with only 2 nodes would finish half-way across the x-axis and not at x-max (in my non-toy sankey diagram this is left-aligned, however, I can't work out the difference)
• Remove the hovertext on nodes only (not on links). I tried various combinations of "label", "text", "value", "percent", "name" joined with a "+" OR "all" or "none" or "skip", but none of this seemed to make a difference.
• Take care of drop-off using NA for example, I don't want to see the link from SA to Drop (blue node), but do want to see the green bar at x=-1 to show one person went to SA on their first holiday and hasn't had another holiday. (If I left source=SA and target=NA, the chart was blank). My proposed work-around will be to otherwise colour DROP Node and SA-DROP links to white...

Have annotated the image with desired changes in blue.

require(dplyr); require(plotly); require(RColorBrewer); require(stringr)

# Summarise flow data
dat <- data.frame(customer = c(1, 1, 1, 2, 2, 2, 2, 3, 3, 4, 4, 5),
              holiday_loc = c("SA", "SA", "AB", "SA", "SA", "SA", "SA", "AB", "AB", "SA", "SA", "SA")) %>%
  group_by(customer) %>%
          mutate(holiday_num = seq_along(customer), 
                 source=paste0(holiday_loc, '_', holiday_num), 
                 target = lead(source),
                 last_hol = ifelse(holiday_num == n(), 'Y', 'N')) %>%
  filter(last_hol== 'N'| holiday_num == 1) %>%
  select(-last_hol)

 sank_links <-  dat %>%
   group_by(source, target) %>%
   summarise(n=n()) %>%
   mutate(target=ifelse(is.na(target), "DROP", target)) # is there another option here?

# obtain colours for nodes
f <- function(pal) brewer.pal(brewer.pal.info[pal, "maxcolors"], pal)
cols <- f("Set1")

# set up nodes
sank_nodes <- data.frame(
                      name = factor(sort(unique(c(as.character(sank_links$source), 
                                   as.character(sank_links$target)))))
                      ) %>%    
                        mutate(label=sub("_[0-9]$", "", name), 
                              # for some unknown reason, plotly allows only three labels to be the same
                              label_pad=sub("_[1-3]$", "", name),
                              label_pad=sub("_[4-6]$", " ", label_pad)) %>%
                        arrange(label) %>%
                        mutate(color = cols[cumsum(1-duplicated(label))])

# update links to get index of node and name (without holiday_num)
sank_links <- sank_links %>%
          mutate(source_num = match(source, sank_nodes$name) -1 , 
                 source_name = str_replace(source, "_[0-9]$", ""),
                 target_num = match(target, sank_nodes$name) - 1,
                 target_name = str_replace(target, "_[0-9]$", ""))


# diagram
p <- plot_ly(
  type = "sankey",
  domain = c(
    x =  c(0,1),
    y =  c(0,1)
  ),
  orientation = "h",
  valueformat = ".0f",
  valuesuffix = "Customers",
  arrangement="fixed",


  node = list(
    label = sank_nodes$label_pad,
    color = sank_nodes$color,
    pad = 15,
    thickness = 15,
    line = list(
      color = "black",
      width = 0.5
    )
  ),

  link = list(
    source = sank_links$source_num,
    target = sank_links$target_num,
    value =  sank_links$n
  )
) %>% 
  layout(
    title = "",
    font = list(
      size = 10
    ),
    xaxis = list(showgrid = F, zeroline = F),
    yaxis = list(showgrid = F, zeroline = F)
  )

p

EDIT: I initially didn't how to to label x-axis with breaks corresponding to nodes and provide title to x-axis; code is as follows:

    %>% 
  layout(
    title = "",
    font = list(
      size = 10
    ),
    xaxis = list(showgrid = F, zeroline = F, title="Holiday Number", tickvals=-1:4, ticktext=1:6),
    yaxis = list(showgrid = F, zeroline = F, showticklabels=FALSE)
  )

Source: https://plot.ly/r/reference/#layout-xaxis-tickformat

Answer 1

You can manually override the nodes position (all of them or only the ones you want).

You can do it in the nodes list, adding a vector for x axis and a vector for y axis with the positions of the nodes that you want to change. If you want to keep a node in the same position, just add NA to that vector position.

node = list(
    label = sank_nodes$label_pad,
    color = sank_nodes$color,
    pad = 15,
    thickness = 15,
    line = list(
      color = "black",
      width = 0.5
    ), 
    x = c(NA, 0.35, 0.65, NA, NA, NA, NA, NA),
    y = c(NA, 0.10, 0.42, NA, NA, NA, NA, NA)
  )

Answer 2

Actually, this is quite possible.

import plotly.graph_objects as go

fig = go.Figure(go.Sankey(
    arrangement = "snap",
    node = {
        "label": ["A", "B", "C", "D", "E", "F"],
        "x": [0.2, 0.1, 0.5, 0.7, 0.3, 0.5],
        "y": [0.7, 0.5, 0.2, 0.4, 0.2, 0.3],
        'pad':10},  # 10 Pixels
    link = {
        "source": [0, 0, 1, 2, 5, 4, 3, 5],
        "target": [5, 3, 4, 3, 0, 2, 2, 3],
        "value": [1, 2, 1, 1, 1, 1, 1, 2]}))

fig.show()

Code from plotly.com.

Answer 3

You can't change the position of the nodes within Plotly but if you change arrangement from 'fixed' to 'freeform' you can move the nodes around manually anywhere you want them after the chart is rendered. However, this has to be done manually by the user every time the chart is rendered. There is no way to order the nodes within Plotly script at the moment.