在 R 中使用 plotly 或 highcharts 嵌套圆环图

Question

library(webr)
data <- as.data.frame(Titanic)  

pd = data %>% 
  group_by(Class, Survived) %>% 
  summarise(n = sum(Freq))

PieDonut(pd,  aes(Class, Survived, count = n),
         labelposition = 0,
         r0 = 0.5,
         r1 = 0.95,
         title = "Titanic: Survial by Class")

This donut fig was made With webr package, is it possible to polt such an interactive fig with plotly or highchart .

Answer 1

1. The plotly way

The input data format for plotly 's sunburst option is confusing (ie a bit of a pain); the relevant example in the documentation is Sunburst with Repeated Labels .

The key is to

1. transform the data in the following way by adding marginal sums (both total and group-level, see my in-line comments), and
2. add a unique ids column.

library(tidyverse)
df <- bind_rows(
    # Full total
    pd %>% summarise(labels = "Total", n = sum(n)),
    # "Class"-level totals
    pd %>% 
        group_by(labels = Class) %>% 
        summarise(
            n = sum(n), 
            parents = "Total",
            .groups = "drop"),
    # Individual Class+Survived-level numbers
    pd %>% 
        rename(parents = Class, labels = Survived) %>%
        mutate(parents = paste("Total", parents, sep = " - "))) %>%
    # Add unique ids column
    mutate(ids = if_else(
        is.na(parents), labels, paste(parents, labels, sep = " - ")))
## A tibble: 13 × 4
#   labels     n parents      ids               
#   <chr>  <dbl> <chr>        <chr>             
# 1 Total   2201 NA           Total             
# 2 1st      325 Total        Total - 1st       
# 3 2nd      285 Total        Total - 2nd       
# 4 3rd      706 Total        Total - 3rd       
# 5 Crew     885 Total        Total - Crew      
# 6 No       122 Total - 1st  Total - 1st - No  
# 7 Yes      203 Total - 1st  Total - 1st - Yes 
# 8 No       167 Total - 2nd  Total - 2nd - No  
# 9 Yes      118 Total - 2nd  Total - 2nd - Yes 
#10 No       528 Total - 3rd  Total - 3rd - No  
#11 Yes      178 Total - 3rd  Total - 3rd - Yes 
#12 No       673 Total - Crew Total - Crew - No 
#13 Yes      212 Total - Crew Total - Crew - Yes

Then

library(plotly)
plot_ly(
    data = df,
    ids = ~ids,
    labels = ~labels,
    parents = ~parents,
    values = ~n,
    type = "sunburst",
    branchvalues = "total")

---

2. The highcharter way

highcharter provides the convenience function data_to_hierarchical() to reformat input data for a sunburst visualisation:

library(tidyverse)
library(highcharter)
pd %>% 
    data_to_hierarchical(c(Class, Survived), n) %>%
    hchart(type = "sunburst")