CODEWITHSUNDEEP
rggplot2dplyr
JasonAizkalns
JasonAizkalns

Reputation: 20463

Subgroup axes ggplot2 similar to Excel PivotChart

I am trying plot Months with Year as subgroups to a chart with ggplot2. That is, something that looks like this:

Pivot Chart

A similar question was answered here, but I am hoping there is a better way that avoids hardcoding the axis labels.

The R code for the data frame is as follows:

set.seed(100)
df = data.frame(       Year = rep(c(rep(2013, 12), rep(2014, 9)), 2)
                ,     Month = rep(rep(month.abb, 2)[1:21], 2)
                , Condition = rep(c("A", "B"), each=21)
                ,     Value = runif(42))

As a bonus, I would appreciate learning how to plot smoothed totals by year without introducing a new variable (if this is possible?). If I use dplyr to summarise and group_by Year and Month, the order of the months is not preserved.

Notice, Month now starts at Apr:

group_by(df, Year, Month) %>% summarise(total = sum(Value)) %>% head
Source: local data frame [6 x 3]
Groups: Year

  Year Month     total
1 2013   Apr 0.4764846
2 2013   Aug 0.9194172
3 2013   Dec 1.2308575
4 2013   Feb 0.7960212
5 2013   Jan 1.0185700
6 2013   Jul 1.6943562

Upvotes: 1

Views: 1319

Answers (1)

baptiste
baptiste

Reputation: 77096

try this,

enter image description here

df$Month <- factor(df$Month, levels=month.abb)

p <- ggplot(df, aes(Month, Value, colour=Condition, group=Condition))+
  facet_grid(.~Year) + geom_line() + theme_minimal() 

library(gtable)
g <- ggplotGrob(p)
g2 <- g[-3,] %>% 
  gtable_add_rows(heights = g$heights[3], nrow(g)-3) %>%
  gtable_add_grob(g[3,], t = nrow(g)-2, l=1, r=ncol(g))

grid.newpage()
grid.draw(g2)

Upvotes: 1

Related Questions