如何在R中绘制相对比例随时间变化

Question

For my research I am trying to create a similar graph based around this graph I found in a piece of literature:

My experiment involved the genre-tagging of 10 different songs. I saved the tags (the words people used to describe seperately). The x-asis should represent all the participants that took part in chronological order. The y-axis should represent how often a word is used in a tag. Consider this sample data:

df <- data.frame(tagid= numeric(0), participantid = numeric(0), tag = character(0))
newRow <-data.frame(tagid=1, participantid=1, tag = "triphop")
df <-rbind(df,newRow)
newRow <-data.frame(tagid=2, participantid=1, tag = "electronic")
df <-rbind(df,newRow)
newRow <-data.frame(tagid=3, participantid=2, tag = "mellow")
df <-rbind(df,newRow)
newRow <-data.frame(tagid=4, participantid=2, tag = "electronic")
df <-rbind(df,newRow)
newRow <-data.frame(tagid=5, participantid=3, tag = "electronic")
df <-rbind(df,newRow)

Tagid 1 and 2 belong to the same participant and should have the same x coordinate. 3 and 4 belong to participant 2 and tagid 5 belongs to participant 3.

For this dataset I'd like to plot a graph like this (excuse the drawing):

The y-axis represents the percentage of participants that have used a specific word to describe this music piece. As 'electronic' is used by all three participants it stays at 100%. 'Triphop' was used by participant 1, but not by participant 2 and 3, decreasing from 100%, to 50%, to 33% at participant 3.

Answer 1

Code is a bit messy, but probably you want something like this ? You need to complete the dataframe so each participantid has rows for all three tag levels. Then, with the cumulative sum of the tag levels and the cumulative sum of participants, you can get the proportion.

df %>%  
  group_by(participantid, tag) %>% 
  summarise(n = n()) %>% 
  complete(tag, nesting(participantid), fill = list(n = 0)) %>%
  group_by(tag) %>% 
  mutate(absolute = cumsum(n)) %>%
  ungroup() %>%
  mutate(id = rep(1:3, each = length(levels(tag)))) %>%
  mutate(proportion = ifelse(absolute / id != 0, absolute / id, NA)) %>%

  ggplot(aes(x = participantid, y = proportion, color = tag)) + geom_line(lwd = 1)