labeling pie chart - strange behaviour in ggplot 2.2.1.

[igreg]

hi all...

I found this example to label a pie chart...

http://stackoverflow.com/questions/24803460/r-ggplot2-add-labels-on-facet-pie-chart

Issue:

### data

year <- c(1,2,1,2,1,2)
prod <- c(1,1,2,2,3,3)
quantity <- c(33,50,33,25,34,25)

df <- data.frame(year, prod, quantity)
rm(year, prod, quantity)



### plot

ibrary(ggplot2)

# center's calculated by hand
centr2 <- c(16, 25, 49, 62.5, 81, 87.5)

ggplot(data=df, aes(x=factor(1), y=quantity, fill=factor(prod))) +
    geom_bar(stat="identity") +
    geom_text(aes(x= factor(1), y=centr2, label = df$quantity), size=10) +
    facet_grid(facets = .~year, labeller = label_value) +
    coord_polar(theta = "y")

The solution from stackoverflow:

#### Solution

library(dplyr)

library(ggplot2)

df <- df %>% group_by(year) %>% mutate(pos = cumsum(quantity)- quantity/2)

ggplot(data=df, aes(x=factor(1), y=quantity, fill=factor(prod))) +

  geom_bar(stat="identity") +

  geom_text(aes(x= factor(1), y=pos, label = quantity), size=10) +  # note y = pos

  facet_grid(facets = .~year, labeller = label_value) +

  coord_polar(theta = "y")

with the latest version of ggplot2 (2.2.1) this solutions doesn't work anymore.

the text labels are in the wrong place... shifted by 180 degrees.

My configuration:

64 bit / windows 7

ggplot2 2.2.1

R version 3.3.2 - 64 bit

any ideas to solve this problem?

thank you!

greg

[ping...@gmail.com]

You have to make it like this,

df <- df %>% group_by(year) %>% mutate(pos = 100-(cumsum(quantity)- quantity/2))

--
--
You received this message because you are subscribed to the ggplot2 mailing list.
Please provide a reproducible example: https://github.com/hadley/devtools/wiki/Reproducibility
 
To post: email ggp...@googlegroups.com
To unsubscribe: email ggplot2+u...@googlegroups.com
More options: http://groups.google.com/group/ggplot2

---
You received this message because you are subscribed to the Google Groups "ggplot2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ggplot2+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[igreg]

thanks for this workaround, jie ping. 

for me, it's still a strange behavior... why did the order of the label position change... it doesn't make sense!

maybe it's the same problem like here

https://groups.google.com/forum/#!topic/ggplot2/buiKAb4RZJw

greg

[san mohan]

When the positions are calculated in dplyr, the data should be ordered in the same order as in the plot. Here is the solution that works: 

year <- c(1,2,1,2,1,2)

prod <- c(1,1,2,2,3,3)

quantity <- c(33,50,33,25,34,25)

df <- data.frame(year, prod, quantity)

rm(year, prod, quantity)

df <- df %>% group_by(year) %>% arrange(year,desc(prod)) %>% mutate(pos = cumsum(quantity)- quantity/2) 

[igreg]

@ san mohan... 

thanks a lot for your solution... this seems to solve the problem in a general way!

have a nice weekend!