How to create a waterfall chart with ggplot2?

Some previous steps are needed to be able to generate that graph, so let's try to generate it while explaining each step.

1. Initialize the data

As you put it in the question, the initialization would be something like this (I have only changed row.namesto nombresto put both in Spanish) and I have added the initial and final balance:

df <- data.frame(
  montos = c(100, 320, 210, -170, -140, 320),
  nombres = c('Saldo Inicial', 'Ingresos', 'Otros ingresos', 'Costos fijos', 'Costos variables', 'Saldo Final')
)

We check that everything is as we wish:

df

## nombres           montos
## Saldo Inicial       100
## Ingresos            320
## Otros ingresos      210
## Costos fijos       -170
## Costos variables   -140
## Saldo Final         320

2. Preserve order

To preserve the order of the data we must convert them nombresto the data type factor.

df$desc <- factor(df$nombres, levels = df$nombres)

3. Add columns

We add a column idand atipo

df$id <- seq_along(df$montos)
df$tipo <- ifelse(df$montos > 0, "in", "out") # Si es positivo será un ingreso y en caso contrario será un egreso
df[df$nombres %in% c("Saldo Inicial", "Saldo Final"), "tipo"] <- "net" # Esto es para el balance inicial y final

Then we add two columns, startand end, this to put the starting and ending value of the amount in each step:

df$end <- cumsum(df$montos)
df$end <- c(head(df$end, -1), 0)
df$start <- c(0, head(df$end, -1))

If I'm not mistaken, you dfshould look like this:

## id  tipo  nombres           start end  montos
## 1   net   Saldo Inicial     0     100  100
## 2   in    Ingresos          100   420  320
## 3   in    Otros ingresos    420   630  210
## 4   out   Costos fijos      630   460 -170
## 5   out   Costos variables  460   320 -140
## 6   net   Saldo Final       320   0    320

4. Generate the graph

Now yes, to generate the graph we do the following:

library(ggplot2)
ggplot(df, aes(nombres, fill = tipo)) + geom_rect(aes(x = nombres,
       xmin = id - 0.45, xmax = id + 0.45, ymin = end,
       ymax = start))

5. Result

Basically that would be it. Some improvements can be made depending on what interests you in the resulting graph.

You can see the full code in this demo .

6. Goodies

This replaces spaces with line breaks to make headings more readable.

strwr <- function(str) gsub(" ", "\n", str)
(p1 <- ggplot(df, aes(fill = tipo)) + geom_rect(aes(x = nombres,
    xmin = id - 0.45, xmax = id + 0.45, ymin = end,
    ymax = start)) + 
    scale_x_discrete("", breaks = levels(df$nombres),
    labels = strwr(levels(df$nombres))) +
    theme(legend.position = "none")
)

As an alternative you can use the waterfall package implemented with lattice ( author's site ).

Example:

library(latticeExtra)
library(waterfall)

df <- data.frame(
  montos = c(420, 210, -170, -140),
  names = c('Ingresos', 'Otros ingresos', 'Costos fijos', 'Costos variables')
)
df$names <- factor(df$names, levels=unique(df$names))
df

##                  montos
## Ingresos            420
## Otros ingresos      210
## Costos fijos       -170
## Costos variables   -140

waterfallchart(montos ~ names, data=df, main="Waterfall graph")

Result In: R-Fiddle