Can I use geom_text or similar to add group by sum rather than amending original data frame?

Question

Here's a bar plot:

ggplot(filtered_funnel, aes(x = reorder(Funnel, -Sessions), y = Sessions)) +
      geom_bar(stat = "identity", fill = "#008080", alpha = 0.6) +
      xlab("Step") +
      ylab("Events") +
      scale_y_continuous(labels = function(l) {l = l / 1000; paste0(l, "K")}) +
      geom_text(aes(label = Sessions, group = Channel), color = "white")

It looks like this (note the text labels in white):

This is because the data df filtered_funnel is actually split out by a field "Channel". I need the bar to be based on the grouped sum of e.g. sessions rather than each individual channel sessions.

Here's a glimpse of the source data:

> glimpse(filtered_funnel)
Observations: 108
Variables: 4
$ Channel  <chr> "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", ...
$ Promo    <chr> "apples", "apples", "apples", "banannas", "banannas", "banannas", "carrots", "carrots", "carrots", "none", "none", "none...
$ Funnel   <chr> "Checkout", "ShippingDetails", "Transactions", "Checkout", "ShippingDetails", "Transactions", "Checkout", "ShippingDetai...
$ Sessions <dbl> 3993, 6332, 2224, 1237, 1962, 689, 2234, 3543, 1244, 42378, 4672, 28120, 87187, 7408, 2602, 611, 969, 340, 4462, 7280, 2...


filtered_funnel
Source: local data frame [108 x 4]
Groups: Channel, Promo [?]

   Channel    Promo          Funnel Sessions
     <chr>    <chr>           <chr>    <dbl>
1   Direct   apples        Checkout     3993
2   Direct   apples ShippingDetails     6332
3   Direct   apples    Transactions     2224
4   Direct banannas        Checkout     1237
5   Direct banannas ShippingDetails     1962
6   Direct banannas    Transactions      689
7   Direct  carrots        Checkout     2234
8   Direct  carrots ShippingDetails     3543
9   Direct  carrots    Transactions     1244
10  Direct     none       AddToCart    42378
# ... with 98 more rows

It looks like ggplot is adding the individual values of each component rather than the sum e.g. just for the first bar in the image (Sessions step)

> filtered_funnel %>% filter(Funnel == "Sessions")
Source: local data frame [6 x 4]
Groups: Channel, Promo [6]

   Channel Promo   Funnel Sessions
     <chr> <chr>    <chr>    <dbl>
1   Direct  none Sessions    87187
2    Email  none Sessions   110035
3 Facebook  none Sessions    79734
4  Organic  none Sessions    80768
5      SEM  none Sessions    94610
6  Youtube  none Sessions    66681

I can see the value 110035 on the image and in the table. What I actually want ggplot to do is add the sum total for sessions.

Since this is being done within a Shiny app I'm trying to avoid medling with the source data since I'm using filter box inputs to filter the data frame. Also, I saw other SO answers that seemed to suggest what I wnat is possible, I just couldn't get it to work e.g.

draw the sum value above the stacked bar in ggplot2

How can I get ggplot to add the sum total for each bar? Can I do this using ggplot grouping and summing rather than altering the source data I feed into ggplot in the aes?

---- dput of data--- Following a comment, here's the data (which has been randomly generated anyway so no NDA issues)

> dput(filtered_funnel)
structure(list(Channel = c("Direct", "Direct", "Direct", "Direct", 
"Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", 
"Direct", "Direct", "Direct", "Direct", "Direct", "Direct", "Direct", 
"Email", "Email", "Email", "Email", "Email", "Email", "Email", 
"Email", "Email", "Email", "Email", "Email", "Email", "Email", 
"Email", "Email", "Email", "Email", "Facebook", "Facebook", "Facebook", 
"Facebook", "Facebook", "Facebook", "Facebook", "Facebook", "Facebook", 
"Facebook", "Facebook", "Facebook", "Facebook", "Facebook", "Facebook", 
"Facebook", "Facebook", "Facebook", "Organic", "Organic", "Organic", 
"Organic", "Organic", "Organic", "Organic", "Organic", "Organic", 
"Organic", "Organic", "Organic", "Organic", "Organic", "Organic", 
"Organic", "Organic", "Organic", "SEM", "SEM", "SEM", "SEM", 
"SEM", "SEM", "SEM", "SEM", "SEM", "SEM", "SEM", "SEM", "SEM", 
"SEM", "SEM", "SEM", "SEM", "SEM", "Youtube", "Youtube", "Youtube", 
"Youtube", "Youtube", "Youtube", "Youtube", "Youtube", "Youtube", 
"Youtube", "Youtube", "Youtube", "Youtube", "Youtube", "Youtube", 
"Youtube", "Youtube", "Youtube"), Promo = c("apples", "apples", 
"apples", "banannas", "banannas", "banannas", "carrots", "carrots", 
"carrots", "none", "none", "none", "none", "none", "none", "pears", 
"pears", "pears", "apples", "apples", "apples", "banannas", "banannas", 
"banannas", "carrots", "carrots", "carrots", "none", "none", 
"none", "none", "none", "none", "pears", "pears", "pears", "apples", 
"apples", "apples", "banannas", "banannas", "banannas", "carrots", 
"carrots", "carrots", "none", "none", "none", "none", "none", 
"none", "pears", "pears", "pears", "apples", "apples", "apples", 
"banannas", "banannas", "banannas", "carrots", "carrots", "carrots", 
"none", "none", "none", "none", "none", "none", "pears", "pears", 
"pears", "apples", "apples", "apples", "banannas", "banannas", 
"banannas", "carrots", "carrots", "carrots", "none", "none", 
"none", "none", "none", "none", "pears", "pears", "pears", "apples", 
"apples", "apples", "banannas", "banannas", "banannas", "carrots", 
"carrots", "carrots", "none", "none", "none", "none", "none", 
"none", "pears", "pears", "pears"), Funnel = c("Checkout", "ShippingDetails", 
"Transactions", "Checkout", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "AddToCart", "Checkout", 
"Registrations", "Sessions", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "Checkout", "ShippingDetails", 
"Transactions", "Checkout", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "AddToCart", "Checkout", 
"Registrations", "Sessions", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "Checkout", "ShippingDetails", 
"Transactions", "Checkout", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "AddToCart", "Checkout", 
"Registrations", "Sessions", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "Checkout", "ShippingDetails", 
"Transactions", "Checkout", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "AddToCart", "Checkout", 
"Registrations", "Sessions", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "Checkout", "ShippingDetails", 
"Transactions", "Checkout", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "AddToCart", "Checkout", 
"Registrations", "Sessions", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "Checkout", "ShippingDetails", 
"Transactions", "Checkout", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions", "AddToCart", "Checkout", 
"Registrations", "Sessions", "ShippingDetails", "Transactions", 
"Checkout", "ShippingDetails", "Transactions"), Sessions = c(3993, 
6332, 2224, 1237, 1962, 689, 2234, 3543, 1244, 42378, 4672, 28120, 
87187, 7408, 2602, 611, 969, 340, 4462, 7280, 2304, 549, 896, 
283, 2094, 3417, 1081, 42251, 5666, 29094, 110035, 9244, 2926, 
256, 418, 132, 129, 191, 85, 3078, 4557, 2039, 120, 178, 79, 
13977, 90, 9727, 79734, 134, 59, 1142, 1691, 756, 3125, 4655, 
1985, 1724, 2568, 1095, 3109, 4631, 1975, 34756, 2864, 23453, 
80768, 4266, 1819, 249, 371, 158, 1839, 2661, 1223, 1543, 2232, 
1026, 2007, 2904, 1335, 24090, 1792, 15272, 94610, 2593, 1192, 
479, 693, 318, 800, 1245, 522, 1734, 2698, 1132, 930, 1447, 607, 
22349, 1436, 14478, 66681, 2235, 937, 1579, 2457, 1031)), class = c("grouped_df", 
"tbl_df", "tbl", "data.frame"), row.names = c(NA, -108L), drop = TRUE, .Names = c("Channel", "Promo", 
"Funnel", "Sessions"), indices = list(0:2, 3:5, 6:8, 9:14, 15:17, 
    18:20, 21:23, 24:26, 27:32, 33:35, 36:38, 39:41, 42:44, 45:50, 
    51:53, 54:56, 57:59, 60:62, 63:68, 69:71, 72:74, 75:77, 78:80, 
    81:86, 87:89, 90:92, 93:95, 96:98, 99:104, 105:107), group_sizes = c(3L, 
3L, 3L, 6L, 3L, 3L, 3L, 3L, 6L, 3L, 3L, 3L, 3L, 6L, 3L, 3L, 3L, 
3L, 6L, 3L, 3L, 3L, 3L, 6L, 3L, 3L, 3L, 3L, 6L, 3L), biggest_group_size = 6L, labels = structure(list(
    Channel = c("Direct", "Direct", "Direct", "Direct", "Direct", 
    "Email", "Email", "Email", "Email", "Email", "Facebook", 
    "Facebook", "Facebook", "Facebook", "Facebook", "Organic", 
    "Organic", "Organic", "Organic", "Organic", "SEM", "SEM", 
    "SEM", "SEM", "SEM", "Youtube", "Youtube", "Youtube", "Youtube", 
    "Youtube"), Promo = c("apples", "banannas", "carrots", "none", 
    "pears", "apples", "banannas", "carrots", "none", "pears", 
    "apples", "banannas", "carrots", "none", "pears", "apples", 
    "banannas", "carrots", "none", "pears", "apples", "banannas", 
    "carrots", "none", "pears", "apples", "banannas", "carrots", 
    "none", "pears")), class = "data.frame", row.names = c(NA, 
-30L), drop = TRUE, .Names = c("Channel", 
"Promo")))

Use `dput(filtered_funnel)` to share your data – GGamba Jun 22 '17 at 06:26 — GGamba, Jun 22 '17 at 06:26
Hi added this now, any help appreciated – Doug Fir Jun 22 '17 at 06:29 — Doug Fir, Jun 22 '17 at 06:29

Axeman · Accepted Answer · 2019-10-31T19:17:08.610

10

You can use stat_summary to calculate the sums, and use that for both the y position and the label:

ggplot(filtered_funnel, aes(x = reorder(Funnel, -Sessions), y = Sessions)) +
  geom_bar(stat = "identity", fill = "#008080", alpha = 0.6) +
  stat_summary(
    aes(label = stat(y)), fun.y = 'sum', geom = 'text', col = 'white', vjust = 1.5
  ) +
  scale_y_continuous(labels = function(l) {l = l / 1000; paste0(l, "K")})  +
  xlab("Step") + ylab("Events")

edited Oct 31 '19 at 19:17

answered Jun 22 '17 at 06:42

Axeman

32,068
8
81
94

Thank you very much (again)! This is great. May I ask some follow ups? What is ..y.. ? Why did you choose (how did you know to choose) `stat_summary()` over `geom_text()`? – Doug Fir Jun 22 '17 at 06:47
`..y..` is the calculated statistic. Each stat (except `identity`) calculates some values before plotting, and you can access those values by their names surrounded with `..`. You can define layers by their `stats` or `geoms`, it usually doesn't matter, you could use `geom_text(aes(label = ..y..), fun.y = 'sum', stat = 'summary', col = 'white', vjust = 1.5)` as well. – Axeman Jun 22 '17 at 06:55
Thanks this really helps. "..y.. is the calculated statistic". The calculated statistics of what? `fun.y = "sum"`? What else could I have added to label (even something nonsensical) just so that it really clicks? (Hope you don't mind the follow ups, I just really value your help here while I have the opportunity) – Doug Fir Jun 22 '17 at 07:01
`stat_summary` summarizes `y` for each `x`. So it calculates a new `y`. You can access this new `y` with `..y..`. If you use `..x..` you get numbers 1 to 6. For other stats you get access to other calculated values. For example, if you look at the [docs page](http://ggplot2.tidyverse.org/reference/geom_histogram.html) for `stat_bin` there is a section "Computed variables" with several options, such as `ndensity`. You can access this variable with `..ndensity..` when using `stat_bin` as your `stat` (either directly or in a `geom`). – Axeman Jun 22 '17 at 07:11
"stat_summary summarizes y for each x". Clicked. Thanks a lot! – Doug Fir Jun 22 '17 at 07:13

Can I use geom_text or similar to add group by sum rather than amending original data frame?

1 Answers1

Linked