Color points by date in ggplot2

问题

Hi all: I am struggling to color points by date in ggplot2. There are two outcomes that would work for me here. 1) colour the points by the variable recent_elections and just add straight lines denoting the date of the most recent election for each point. The current code does that. 2) preferably, but harder, just add the lines, coloured differently for each election, showing a legend that printed the date of the most recent federal election.

My current data and attempt is below.

library(dplyr)
library(tidyr)
library(ggplot2)

members <- structure(list(date = structure(c(6209, 6574, 7305, 14984, 15339, 
15341, 17169, 17174), class = "Date"), members = c(180835, 193225, 
200010, 86545, 95000, 128351, 41000, 124000), population = c(26449000, 
26798000, 27512000, 33476688, 33476688, 33476688, 35151728, 35151728
), votes_previous_election = c(2359915, 2685263, 2685263, 4508474, 
4508474, 4508474, 3470350, 3470350), vote_percent = c(18.8, 20.4, 
20.4, 30.6, 30.6, 30.6, 19.7, 19.7), seats_previous_election = c(32, 
43, 43, 103, 103, 103, 44, 44), recent_election = structure(c(5360, 
6899, 6899, 15096, 15096, 15096, 16727, 16727), class = "Date")), .Names = 
c("date", 
"members", "population", "votes_previous_election", "vote_percent", 
"seats_previous_election", "recent_election"), class = "data.frame", 
row.names = c(NA, 
-8L))

members %>% 
  select(population, votes_previous_election, seats_previous_election, members, 
         date, recent_election) %>% 
  mutate(., members_per_capita=members/population, 
         members_votes=members/votes_previous_election, 
         members_seats=members/seats_previous_election) %>% 
  gather(Variable, Value, c(members_per_capita,members_votes, 
                            members_seats))%>% 
  ggplot(., aes(x=date, y=Value, 
                group=recent_election))+
  geom_point(aes(fill=recent_election))+
  facet_wrap(~Variable, scales='free')+
  geom_vline(data=members, aes(xintercept=as.numeric(recent_election), col='red'), show.legend=F)

回答1:

members %>% 
  select(population, votes_previous_election, seats_previous_election, members, 
         date, recent_election) %>% 
  mutate(., members_per_capita=members/population, 
         members_votes=members/votes_previous_election, 
         members_seats=members/seats_previous_election) %>% 
  gather(Variable, Value, c(members_per_capita,members_votes, 
                            members_seats))%>% 
  ggplot(., aes(x=date, y=Value, 
                group=recent_election))+
  geom_point()+
  geom_vline(data=members, aes(xintercept=as.numeric(recent_election), col=factor(recent_election)), show.legend=T)+
  facet_wrap(~Variable, scales='free') +
  scale_color_discrete(name = "Recent Election") + xlim(as.Date("1984-01-01"), NA)

I changed the col="red" in geom_vline to col=factor(recent_election) so that the vertical lines are colored by recent_election. The factor() makes sure that recent_election is treated as discrete instead of continuous. scale_color_discrete sets the legend title. Note that the election date "1984-09-04" is going out of the x range of your points, so I added a xlim(as.Date("1984-01-01"), NA) to also include that election date. NA sets the upper limit automatically.

来源：https://stackoverflow.com/questions/45965155/color-points-by-date-in-ggplot2

标签

date

ggplot2