Change plotting order of an area chart in R

Question

I am using the following code to try to plot a graph

ggplot(final2, aes(x = final2$a, y =final2$somaacumulada,fill=final2$Country)) +
  theme_bw()+
  geom_area(position = position_dodge(width = 0.8))

This is a part of data frame

"Rated Power","Country","Commissioned","d","m","a","somaacumulada"
1.14,"Resto do mundo","01.01.1985",1,1,1985,17451.5
1.28,"Japan","01.01.1995",1,1,1995,16969.78
1.7,"Resto do mundo","01.01.1966",1,1,1966,2542.3
2,"China","01.11.1992",1,11,1992,485
3.5,"Germany","01.01.1973",1,1,1973,1346.7
3.6,"Spain","01.01.1955",1,1,1955,10.8
4.6,"Resto do mundo","01.01.1957",1,1,1957,238.6
7,"Austria","01.01.1969",1,1,1969,352
7.2,"Spain","01.01.1929",1,1,1929,7.2
8.5,"United States","01.01.1954",1,1,1954,37.5
11,"France","01.01.1951",1,1,1951,11
11.3,"Spain","27.06.2014",27,6,2014,6758.7
12,"India","01.01.1976",1,1,1976,162
13.5,"Japan","01.01.1961",1,1,1961,13.5
14,"Spain","01.01.1966",1,1,1966,339.8
15.9,"Resto do mundo","01.01.1969",1,1,1969,2758.2
17,"Resto do mundo","01.01.1986",1,1,1986,18791.5
22,"China","01.11.1973",1,11,1973,63
24,"Resto do mundo","01.01.1953",1,1,1953,234
25.2,"United States","01.01.1973",1,1,1973,6518.7
28,"United States","01.01.1972",1,1,1972,5277.5
29,"United States","01.01.1929",1,1,1929,29
30,"Resto do mundo","01.01.1937",1,1,1937,165
30.8,"Resto do mundo","01.01.1989",1,1,1989,20912.3
35,"Germany","01.01.1959",1,1,1959,404
36.5,"Italy","01.01.1965",1,1,1965,489.7
37,"Resto do mundo","31.12.2016",31,12,2016,33017.7
40,"Resto do mundo","01.01.1974",1,1,1974,5482.2
40,"United States","14.09.2012",14,9,2012,22560.7
41,"China","01.05.1968",1,5,1968,41
42,"Italy","01.01.1968",1,1,1968,646.7
45,"Resto do mundo","01.01.1944",1,1,1944,210
45,"United States","01.01.1994",1,1,1994,21425.7
48,"France","01.01.1957",1,1,1957,59
49,"Germany","01.01.1983",1,1,1983,3999.7
49.2,"Germany","01.01.1960",1,1,1960,453.2
50,"Austria","01.01.2011",1,1,2011,3200
50,"United States","01.08.1971",1,8,1971,5249.5
53,"Resto do mundo","01.01.1990",1,1,1990,20965.3
54,"Spain","01.01.1982",1,1,1982,2561.3
58,"United States","01.01.1984",1,1,1984,15300.9

However I am not able to make the order of the areas right, the larger areas that are in front of the graph and hide the smaller ones that are behind, someone would know a way to solve this?

Another question I have would be to limit the values on the X axis to have a range that better represents the graph.

EDIT --

I made the following editions:

final2$Country <- factor(final2$Country, levels = c('Resto do mundo','United States','Japan','China','Italy','Germany','Spain','France','India','Austria'))

ggplot(final2, aes(x = Commissioned, y =somaacumulada)) +
  theme_bw()+
  geom_area(position = position_dodge(width = 0.8),alpha=1, aes(fill=Country)) +
  scale_x_date(breaks= "10 years", labels = date_format("%Y"))+
  scale_y_continuous(breaks=seq(0,37000,1000)) +
  xlab('Ano') +ylab('Potência Instalada [MW]')

And I got the following result

however the result is not suitable, still has some hidden values and, as I am using accumulated sum, the values of any country should not decrease from one year to another, or increases or remains constant.

Answer 1

Here are two solutions to the problem. Both order the countries of the largest somaacumulada for the smallest but do it in different ways and the final orders are different. In both solutions Country becomes a class column "factor", different levels depending on the order of somaacumulada.

1. Load the package ggplot2 and turn dates into true dates.

library(ggplot2)

final2$Commissioned <- as.Date(final2$Commissioned, "%d.%m.%Y")

2. First solution.

This solution orders countries by date of Commissioned and somaacumulada.

1. Create an index i;
2. Create a level vector for the factor Country by order of i.
3. Create the factor with these levels.

The result is that we have the largest accumulated sums to last date in the background, then the big seconds, etc. The smaller areas are in the foreground but only those that reach the last day.

i <- with(final2, order(Commissioned, somaacumulada, decreasing = TRUE))
lvls <- unique(final2$Country[i])
final2$Country <- factor(final2$Country, levels = lvls)

ggplot(final2, aes(x = a, y = somaacumulada, fill = Country)) +
  geom_area(position = position_dodge(width = 0.8)) +
  theme_bw()

3. Second solution.

Now the countries are ordered by somaacumulada without taking into account the dates. The graphic instructions are exactly the same.

m <- with(final2, tapply(somaacumulada, Country, max))
m <- sort(m, decreasing = TRUE)
lvls <- unique(names(m))
final2$Country <- factor(final2$Country, levels = lvls)

ggplot(final2, aes(x = a, y = somaacumulada, fill = Country)) +
  geom_area(position = position_dodge(width = 0.8)) +
  theme_bw()

Editing

To answer the question under review, how to add one line per country for the year 2019 with the same data as the last row for that country, if there is no data yet for 2019, the following code is a solution in R base.

ano <- 2019
d.ano <- as.Date(paste(ano, 1, 1, sep = '-'))
sp <- split(final2, final2$Country)

sp <- lapply(sp, function(X){
  n <- nrow(X)
  if(X[n, 'a'] < ano){
    new <- X[n,, drop = FALSE]
    new$Commissioned <- d.ano
    new$d <- 1
    new$m <- 1
    new$a <- ano
  }
  rbind(X, new)
})
novo <- do.call(rbind, sp)
row.names(novo) <- gsub('[^[:digit:]]+', '', row.names(novo))
novo <- novo[order(as.integer(row.names(novo))), ]
novo