How do I make a proportional area chart?

Question

I am currently trying to get a proportional data chart similar to the one found on this page (http://www.improving-visualisation.org/vis/id=148).

The data I am working from is extracted from a CSV of the number of college graduates any given year stratified by sex and type of course from 1993 to 2003 inclusive. I have split the csv into "df_list"- the 1993 can be seen here.

df_list[1]$1993

   year     sex                            type_of_course no_of_graduates
1  1993   Males                                 Education              na
2  1993   Males                              Applied Arts              na
3  1993   Males              Humanities & Social Sciences             481
4  1993   Males                        Mass Communication              na
5  1993   Males                               Accountancy             295
6  1993   Males                 Business & Administration             282
7  1993   Males                                       Law              92
8  1993   Males Natural, Physical & Mathematical Sciences             404
9  1993   Males                                  Medicine              95
10 1993   Males                                 Dentistry              14
11 1993   Males                           Health Sciences              10
12 1993   Males                    Information Technology             264
13 1993   Males                   Architecture & Building             132
14 1993   Males                      Engineering Sciences            1496
15 1993   Males                                  Services              na
16 1993 Females                                 Education              na
17 1993 Females                              Applied Arts              na
18 1993 Females              Humanities & Social Sciences            1173
19 1993 Females                        Mass Communication              na
20 1993 Females                               Accountancy             396
21 1993 Females                 Business & Administration             708
22 1993 Females                                       Law              93
23 1993 Females Natural, Physical & Mathematical Sciences             588
24 1993 Females                                  Medicine              61
25 1993 Females                                 Dentistry              11
26 1993 Females                           Health Sciences              40
27 1993 Females                    Information Technology             215
28 1993 Females                   Architecture & Building             144
29 1993 Females                      Engineering Sciences             254
30 1993 Females                                  Services              na

I understand the next step will to make individual proportional bar graphs for each year with respect to each course- how exactly to I go about that? I am currently stuck trying to merge males and females into a single row.

Answer 1

Not having the data to work with, I generated a small sample which might be used by others and improved :

library(dplyr)
library(ggplot2)
set.seed(1)
year <- rep("1993", 20)
sex <- rep(c("M","F"), each = 10)
course <- rep(letters[1:10], 2)
num <- round(runif(20, min = 49, max = 120))
df <- data.frame(year, sex, course, num)
df <- tbl_df(df)
df
Source: local data frame [20 x 4]

     year    sex course   num
   (fctr) (fctr) (fctr) (dbl)
1    1993      M      a    68
2    1993      M      b    75
3    1993      M      c    90
4    1993      M      d   113
5    1993      M      e    63
6    1993      M      f   113
7    1993      M      g   116
8    1993      M      h    96
9    1993      M      i    94
10   1993      M      j    53
11   1993      F      a    64
12   1993      F      b    62
13   1993      F      c    98
14   1993      F      d    76
15   1993      F      e   104
16   1993      F      f    84
17   1993      F      g   100
18   1993      F      h   119
19   1993      F      i    76
20   1993      F      j   104

Note that I used letters for the course names. Now I will make a labeling dataset:

mylab <- data.frame(x = factor(df$course), 
         y = rep(c(0.3, 0.8), each = 10), l = df$num)

Finally to the plot:

ggplot(data = df, aes(x = factor(course), y = num, fill=factor(sex))) +
     geom_bar(stat = "identity", position = "fill") + 
     geom_text(data = mylab, aes(x = x, y = y, label = l))

The output is:

Of course now one can play with the labels and title.
Hope it helps.

Answer 2

Although the final question is to obtain a graph, seems more a question of handling data.frame.

A toy-dataset for a reproducible example:

year <- c(rep("1993",6), rep("1994",6))
sex <- c(rep("males",3), rep("females", 3))
course <- c(rep(letters[1:3],4))
number <- 1:12*10
data <- data.frame(cbind(year, sex, course, number))
data$number <- as.numeric(data$number)
data$number[1] <- NA

And unify the number variable regardless of the sex variable.

library(dplyr)
df <- data %>% group_by(year, course) %>% summarise(total=sum(number, na.rm=TRUE))     
library(plyr)
df_2 <- ddply(df, .(year), transform, label_y=cumsum(total))

footnote(1)

Obtaining the desired chart

library(ggplot2)
ggplot(df_2, aes(x=year, y=total, fill=course)) + geom_bar(stat="identity") + 
      geom_text(aes(y=label_y ,label=total), vjust=3, colour="white")  

(1) The loading of packages and play dplyr is very sensitive. There are functions that are modified and to repeat the process a second time have to leave R. I have not found a better way.

Answer 3

Because the data isn't reproducible this is untested. But I think that something like this would work:

library(ggplot2)

ggplot(dflist, aes(x=sex, y=no_of_graduates, fill=type_of_course)) +
   geom_bar(stat="identity")

Or a trivial iris example:

library(dplyr)

iris %>%
  mutate(DummyXVar="DummyX") %>%
  ggplot(aes(y=Petal.Width, x=DummyXVar,fill=Species)) +
  geom_bar(stat="identity")

HTH