How to add AUC to a multiple ROC graph with pROC's ggroc

Question

I have a list of elements of the class "roc" (l_rocs) which I want to plot with ggroc from the package pROC

library("ggplot2")
library("pROC")

#inside a bigger loop
l_rocs[[names[[i]]]] <- roc(predictor=gbm.probs$Yes,
                response=testing$Attrition,
                levels=levels(testing$Attrition))

#loop end

ggroc(l_rocs) +
  labs(color='Sampling Method'))

I now want to add the AUC for each curve. The best would be right inside the legend but I can not find a way to do it as the given element is a list.

Any advice?

Answer 1

Here is a solution using modified legend labels

I used some example data since no data for a reproducible example was added.

#library
library(pROC)
library(ggplot2)
library(tidyverse)


# example data
roc.list <- roc(outcome ~ s100b + ndka + wfns, data = aSAH)
#> Setting levels: control = Good, case = Poor
#> Setting direction: controls < cases
#> Setting levels: control = Good, case = Poor
#> Setting direction: controls < cases
#> Setting levels: control = Good, case = Poor
#> Setting direction: controls < cases

# extract auc
roc.list %>% 
  map(~tibble(AUC = .x$auc)) %>% 
  bind_rows(.id = "name") -> data.auc

# generate labels labels
data.auc %>% 
  mutate(label_long=paste0(name," , AUC = ",paste(round(AUC,2))),
         label_AUC=paste0("AUC = ",paste(round(AUC,2)))) -> data.labels

# plot on a single plot with AUC in labels
ggroc(roc.list) +
  scale_color_discrete(labels=data.labels$label_long)

If you have multiple ROC curves it might be better to draw a facet plot

# plot a facet plot with AUC within plots
ggroc(roc.list) +
  facet_wrap(~name) +
  
  geom_text(data = data.labels,
          aes(0.5, 1, 
              label = paste(label_AUC)),
          hjust = 1) 

^{Created on 2021-09-30 by the reprex package (v2.0.1)}