Display Counts on a ggplot

Question

I want to display the total number of lengths per year on my graphs. How do I display the total number of lengths (FREQUENCY) per year on my graph? I do NOT want it displayed above each bar. I want the total number for that year.

ggplot(data = OT_LF, aes(x = Length_mm, y = FREQUENCY))+
geom_bar(stat="identity", width = 10, fill = "black")+
theme_bw()+
labs(x = "Length (mm)", y = "Count")+
facet_wrap(~Year, ncol = 1, dir="v")

Here's my data:

dput(head(OT_LF))
structure(list(ID = c(20154113L, 20154113L, 20154113L, 20154113L, 
20154113L, 20154113L), CRUCODE = c(20154L, 20154L, 20154L, 20154L, 
20154L, 20154L), Cruise = c(4L, 4L, 4L, 4L, 4L, 4L), Year = c(2015L, 
2015L, 2015L, 2015L, 2015L, 2015L), Month = c(8L, 8L, 8L, 8L, 
8L, 8L), STRATUM = c(16L, 16L, 16L, 16L, 16L, 16L), TOW = c(16L, 
16L, 16L, 16L, 16L, 16L), STA = c(113L, 113L, 113L, 113L, 113L, 
113L), YRMODA = c(20150821L, 20150821L, 20150821L, 20150821L, 
20150821L, 20150821L), MINOUT = c(20, 20, 20, 20, 20, 20), SPP = c(136L, 
136L, 136L, 136L, 136L, 136L), LENGTH = c(28L, 24L, 29L, 25L, 
23L, 26L), Length.mm. = c(280L, 240L, 290L, 250L, 230L, 260L), 
FREQUENCY = c(2L, 4L, 1L, 3L, 3L, 2L), Length_mm = c(280, 
240, 290, 250, 230, 260)), row.names = 4257:4262, class = "data.frame")

You can pre-compute `cnts <- tapply(OT_LF$FREQUENCY, OT_LF$Year, sum)` and then add a layer `geom_text(aes(x = 280, y = 4, label = cnts))`. — Rui Barradas, Jun 25 '20 at 18:06
I tried what you said, but I get this error 'Error: Aesthetics must be either length 1 or the same as the data (320): label' What does that mean? — Stacy, Jun 25 '20 at 18:13
Try `agg <- aggregate(FREQUENCY ~ Year, OT_LF, sum)` and `geom_text(agg, mapping = aes(x = 280, y = 4, label = cnts))`. — Rui Barradas, Jun 25 '20 at 18:22

score 2 · Answer 1 · answered Jun 25 '20 at 18:29

Here's an approach with the ..PANNEL.. special symbol:

ggplot(data = OT_LF, aes(x = Length_mm, y = FREQUENCY))+
  geom_bar(stat="identity", width = 10, fill = "black")+
  geom_text(aes(label = paste0("Total: ",tapply(OT_LF$FREQUENCY, OT_LF$Year, sum)[..PANEL..]),
                x = diff(range(Length_mm))/2 + min(Length_mm),
                y = max(FREQUENCY) + 1)) + 
  scale_y_continuous(expand = expansion(mult = c(0, .2))) + 
  theme_bw()+
  labs(x = "Length (mm)", y = "Count")+
  facet_wrap(~Year, ncol = 1, dir="v")

Edited Sample Data

OT_LF <- structure(list(ID = c(20154113L, 20154113L, 20154113L, 20154113L, 
20154113L, 20154113L), CRUCODE = c(20154L, 20154L, 20154L, 20154L, 
20154L, 20154L), Cruise = c(4L, 4L, 4L, 4L, 4L, 4L), Year = c(2015L, 
2015L, 2015L, 2015L, 2018L, 2018L), Month = c(8L, 8L, 8L, 8L, 
8L, 8L), STRATUM = c(16L, 16L, 16L, 16L, 16L, 16L), TOW = c(16L, 
16L, 16L, 16L, 16L, 16L), STA = c(113L, 113L, 113L, 113L, 113L, 
113L), YRMODA = c(20150821L, 20150821L, 20150821L, 20150821L, 
20150821L, 20150821L), MINOUT = c(20, 20, 20, 20, 20, 20), SPP = c(136L, 
136L, 136L, 136L, 136L, 136L), LENGTH = c(28L, 24L, 29L, 25L, 
23L, 26L), Length.mm. = c(280L, 240L, 290L, 250L, 230L, 260L), 
FREQUENCY = c(2L, 4L, 1L, 3L, 3L, 2L), Length_mm = c(280, 
240, 290, 250, 230, 260)), row.names = 4257:4262, class = "data.frame")

@Stacy You should accept this answer, it's general purpose enough for the question. And the `x`, `y` coordinates are great. — Rui Barradas, Jun 25 '20 at 18:57
I have one last question, because I'm new to R, how do you do count rather than sum using the tapply function? — Stacy, Jun 25 '20 at 19:13
@Stacy I'm not totally sure what you mean, but maybe `function(x) length(unique(x))`? — Ian Campbell, Jun 25 '20 at 20:02

Duck · Accepted Answer · 2020-06-26T00:44:56.900

1

Try this (Updated):

library(dplyr)
library(ggplot2)

OT_LF %>% left_join(OT_LF %>% group_by(Year) %>% summarise(TT=sum(FREQUENCY,na.rm=T))) %>%
  mutate(NewFacet=paste0(Year,' (N=',TT,')')) -> DF2

ggplot(data = DF2, aes(x = Length_mm, y = FREQUENCY))+
  geom_bar(stat="identity", width = 10, fill = "black")+
  theme_bw()+
  labs(x = "Length (mm)", y = "Count")+
  facet_wrap(~NewFacet, ncol = 1, dir="v")

Updated for histogram:

OT_LF %>% left_join(OT_LF %>% group_by(Year) %>% summarise(TT=n())) %>%
    mutate(NewFacet=paste0(Year,' (N=',TT,')')) -> DF2

ggplot(data = DF2, aes(x = Length_mm))+
    geom_histogram( stat = "bin", position = "stack", binwidth = 3, fill = "black" )+
    theme_bw()+
    labs(x = "Length (mm)", y = "Count")+
    facet_wrap(~NewFacet, ncol = 1, dir="v")

edited Jun 26 '20 at 00:44

answered Jun 25 '20 at 18:05

Duck

39,058
13
42
84

I don't want it displayed per column. I just want it to display the total number for that year. Is there a way to do that? – Stacy Jun 25 '20 at 18:07
@Stacy I have updated the solution. Hoping this can help. – Duck Jun 25 '20 at 19:32
Thank you! This is very helpful. I've tried to repeat this solution for another dataset, but instead of getting sum of FREQUENCY, I need to get count per year. I've adjusted my code to this: DRS_LF %>% left_join(DRS_LF %>% group_by(Year) %>% summarise(TT=count(DRS_LF, Year, na.rm=T))) %>% mutate(NewFacet=paste0(Year,' (N=',TT,')')) -> DF3 but it gives me this error when I try to run my histogram, Error: At least one layer must contain all faceting variables: `NewFacet`. * Plot is missing `NewFacet` * Layer 1 is missing `NewFacet`.......do you know what I'm doing wrong? – Stacy Jun 25 '20 at 21:55
@Stacy Try `OT_LF %>% left_join(OT_LF %>% group_by(Year) %>% summarise(TT=n())) %>%mutate(NewFacet=paste0(Year,' (N=',TT,')')) -> DF2` – Duck Jun 25 '20 at 22:18
@Stacy Or more precisely `DRS_LF %>% left_join(DRS_LF %>% group_by(Year) %>% summarise(TT=n())) %>% mutate(NewFacet=paste0(Year,' (N=',TT,')')) -> DF3` – Duck Jun 25 '20 at 22:21
It did not work. I am also trying to use a geom_histogram instead though. Could that be the issue? Here's my code:ggplot(DRS_LF, aes(x = LENGTH))+ geom_histogram( stat = "bin", position = "stack", binwidth = 3, fill = "black" )+ theme_bw()+ labs(x = "Length (mm)", y = "Count")+ facet_wrap(~Year, ncol = 1, dir="v") – Stacy Jun 25 '20 at 22:38
@Stacy I have included a sketch for histogram in the solution using Ian's data. Hope it can be useful. – Duck Jun 26 '20 at 00:45
Ugh yes! It worked and I realized what I did wrong! I was trying to use my original data file instead of the new one created for the summarized data. It's the little things that get ya! THANK YOU!! – Stacy Jun 26 '20 at 01:53

Display Counts on a ggplot

2 Answers2