Barchart for multiple variables

Question

[enter image description here][1]

I have a table from a questionnaire where 390 people answered the likelihood for an activity as either unlikely, likely or very likely. The table looks as follows (for a total of 390 lines)

Anniversary	Babyshower	Engagement	Memorial
Unlikely	Likely	Very likely	Unlikely
Likely	Unlikely	Likely	Very likely
Very likely	Very likely	Unlikely	Likely

What I would like is a barchart for each columns separated by the occurences of unlikely, likely and very likely. I am able to pull a barchart for one column, however I would like to have a a chart where I can see each individual column in the same graph. Is there a way to do this in ggplot?

I would be grateful for your help.

Thanks.

user_15133590 · Answer

library(ggplot2)

#creating example dataset
dfx <- data.frame(
  Anniversary = c(sample(c("likely", "unlikely", "very_likely"), 20, replace = TRUE)),
  Annual_album = c(sample(c("likely", "unlikely", "very_likely"), 20, replace = TRUE)),
  Babyshower = c(sample(c("likely", "unlikely", "very_likely"), 20, replace = TRUE)),
  Childhoodmemories = c(sample(c("likely", "unlikely", "very_likely"), 20, replace = TRUE)),
  stringsAsFactors = FALSE
)


#creating a function to transform any dataset to a grouped dataset based on dataset variables(here activities)
grouped_df <- function(dataset){
      #creating a dataframe with each response count per activities
      df_table <- as.data.frame(apply(dataset,2, table))  
      
      # creating a vector containing activities
      activity <- rep(colnames(df_table), each = nrow(df_table))
      
      # creating a vector containing responses(likely, unlikely, very_likely)
      response <- rep(rownames(df_table), times = length(colnames(df_table)))
      
      # creating a vector containing number of responses per activities(vote_num)
      vote_num <- NULL    
      for (i in seq_along(colnames(df_table))){           
        name <- paste0(colnames(df_table)[i], "_num")
        name <- assign(name, df_table[,i])
        vote_num <- append(vote_num, name)
      }
      
      
      # creating final dataframe with 3 columns(activities, responses, vote_num)
      df_final <- data.frame(
        activity = activity,
        response = response,
        vote_num = vote_num,
        stringsAsFactors = FALSE
      )
      return(df_final)     #retuning dataframe
}

my_dataset <- grouped_df(dfx)      #assigning the function to a variable

# creating ggplot of grouped barchart
ggplot(my_dataset, aes(fill = response, y = vote_num, x = activity)) +
  geom_bar(position = "dodge", stat = "identity")

It creates a grouped barchart as below:

Did you mean something like this?

In the response of the issues of @Falimus after creating plots, which shows (1,2,3) instead of labels:
Code above worked in my computer with different data types, but I wrote the code below, maybe it solved the remained issue of @Falimus:

#In the ggplot part(last line), write this code instead
ggplot(my_dataset, aes(fill = response, y = vote_num, x = activity)) +
  geom_bar(position = "dodge", stat = "identity") + scale_fill_discrete(labels=c("likely", "unlikely", "very_likely"))   #the orders is from above the legend

user_7021497 · Answer

Are you looking for something like this?

df<-data.frame(Anniversary = c("Unlikely", "Likely", "Very Likely"),
               Babyshower = c("Likely", "Unlikely", "Very Likely"),
               Engagement = c("Very Likely", "Likely", "Unlikely"))
df %>%
  pivot_longer(., cols = c("Anniversary", "Babyshower", "Engagement")) %>%
  group_by(name, value) %>%
  summarise(total = n()) %>%
  ggplot(.) + 
  geom_col(
           mapping = aes(x = name, y = total, fill = value), width = 0.5,
           position = position_dodge(width = 0.6))

Barchart for multiple variables

2 Answers