Color scale adjustment on R

Question

Good afternoon,

I created a bar chart using an indicator to fill the colors of the bars, where the nearest red is bad, and the nearest green is good. It turns out that it adjusted according to the data, so that the number 29 turned green, when it is a very bad value. I would like to manually adjust the range that will adjust the colors. How to do? Thank you!

graf_pecas_tipo <- dados_expandido %>% 
    filter(hierarquia0 == "PEÇAS + ACESSORIOS") %>% 
      group_by(hierarquia3) %>% 
       summarise(Qtd_busc = n(),
                Qtd_vend = sum(ind_venda == "S"),
                Tx_Conv = round(Qtd_vend / Qtd_busc * 100,2),
                Clie_dt = n_distinct(cpf_cliente),
                Preco_Med_vend = round(mean(valor_unitario_sem_acrescimo[ind_venda == "S"]),2),

            Preco_Med_sem_vend = round(mean(valor_unitario_sem_acrescimo[ind_venda == "N"]),2),
            Desc_Med = round(mean(desconto, na.rm = T)),2) %>% 

  arrange(desc(Qtd_busc))%>% 
  top_n(30, Qtd_busc) %>% 
  mutate(desc_produto = reorder(hierarquia3, - Qtd_busc),
         Tx_Conv = as.integer(Tx_Conv))%>% #ordenando as barras do gráfico

  ggplot(aes(x = desc_produto, y = Qtd_busc)) +
  geom_col(aes(fill = Tx_Conv)) +
  geom_text(aes(label = Tx_Conv, vjust = -0.5))+
  scale_fill_distiller(type = "div", palette = "RdYlGn", direction = 1) +
  labs(x = "Tipo de Peça",
       y = "Quantidade de Orçamentos",
       title = "Quantidade de buscas e Taxa de Conversão por Tipo de Peça",
       subtitle = NULL,
       fill = "Taxa de Conversão")+
  theme_bw()+
  theme(axis.text.x=element_text(angle = 45, hjust = 1))


ggplotly(graf_pecas_tipo)

Answer 1

First, I created a database that resembles your:

# Criando conjunto de dados
df <- data.frame(Var1 = 9,
   Var2 = 5,
   Var3 = 7,
   Var4 = 4,
   Var6 = 9,
   Var7 = 8)
long <- reshape2::melt(df)
long
  variable value
1     Var1     9
2     Var2     5
3     Var3     7
4     Var4     4
5     Var6     9
6     Var7     8

One of the ways to find a color palette, and do it automatically is via the package RColorBrewer. When performing the function display.brewer.all() you will see different color palettes that you can use. Another option is to use their website colorbrewer2. In this case, the palette that serves in your case is "Rdylgn":

library(ggplot2)
library(RColorBrewer)
ggplot(long, x = variable, y = value) +
  geom_col(aes(x = variable, y = value, fill = variable)) +
  scale_y_continuous(labels = 0:10, breaks = 0:10) +
  geom_text(aes(label = value, x = variable, y = value)) +
  scale_fill_brewer(palette = "RdYlGn", direction = 1)

If you use direction = -1, colors will appear in reverse order. The limitation of this palette is that it displays at most 11 different colors.

Alternatively, use colorRampPalette and put the number of colors you want in the function you create, this would be the "manual" mode. In this case, I left as the number of lines:

colfunc <- colorRampPalette(c("red", "green"))
ggplot(long, x = variable, y = value) +
  geom_col(aes(x = variable, y = value, fill = variable)) +
  scale_y_continuous(labels = 0:10, breaks = 0:10) +
  scale_fill_manual(values = colfunc(nrow(long)))

Answer 2

Use ..count.. as parameter for function Fill attribute ggplot(), as the example below:

ggplot(df) + geom_bar(aes(year,fill=..count..))

Source: https://stackoverflow.com/a/36827261/10328310