CODEWITHSUNDEEP
rdataframeplotggplot2
GCGM
GCGM

Reputation: 1073

Update graph based on table values in R

I am running a code in R that has several iterations. The result of each one is stored in a table called accatable. As you can see, in this example below there is only the result for the row S2* 

> accatable
              1         2        3         4         5         6        7
S1           NA        NA       NA        NA        NA        NA       NA
S2           NA        NA       NA        NA        NA        NA       NA
S1_S2        NA        NA       NA        NA        NA        NA       NA
S2*    0.737714 0.7083141 0.767515 0.8060774 0.7800401 0.8015116 0.815209
S1_S2*       NA        NA       NA        NA        NA        NA       NA

What I want to create is a graph using ggplot2 showing the evolution. For example, you run the first iteration and get the value for row S2* column 1. Then in the second iteration you get the value for row S2* column 2, etc.

The objective is that after each iteration you plot a graph that will be updated each time to show the evolution.

So far, I have manage to create that graph but only when all the table is completed. Here is the test I have tried. I first create the df and convert it from wide to long format. Then I used ggplot to crate the output

testdf <- replicate(7, sample(0:10,5,rep=TRUE))
colnames(testdf) <- as.character(seq(1,7))
rownames(testdf) <- c("S1", "S2", "S1_S2", "S2*", "S1_S2*")
test <- melt(testdf, id.vars=testdf[[1]])
colnames(test) <- c("Input", "Images", "Acca")
test

test$IMAGES <- as.numeric(as.vector(test$Images))

ggplot (data = test, aes(x=Images, y=Acca, group=Input, colour=Input)) + 
  geom_line(aes(linetype=Input)) + 
  geom_point() + 
  scale_colour_manual(name="Scenario", 
                      values = c("black","black","blue","blue","red","red", 
                                 "darkgreen","darkgreen")) + 
  scale_linetype_manual(name="Scenario",
                        values=c("solid","dashed","solid","dashed","solid", "dashed", 
                                 "solid","dashed","solid","dashed", "solid","dashed")) + 
  theme_minimal() + 
  labs(x="Images", y="Acca",title="test") + 
  theme(plot.title = element_text(hjust = 0.5)) + 
  scale_x_continuous("Images", c(1,2,3,4,5,6,7), c(1,2,3,4,5,6,7))

Any idea how I could adapt the ggplot code to plot the table each time a new value is added?

Upvotes: 0

Views: 67

Answers (1)

Z.Lin
Z.Lin

Reputation: 29095

Here's a tidyverse solution.

In order to illustrate this, I've created a blank data frame the same size as testdf, to be updated iteratively:

testdf <- as.data.frame(testdf)
accatable <- data.frame(`1` = rep(NA, 5), `2` = rep(NA, 5),
                        `3` = rep(NA, 5), `4` = rep(NA, 5),
                        `5` = rep(NA, 5), `6` = rep(NA, 5),
                        `7` = rep(NA, 5),
                        row.names = rownames(testdf))

> accatable
       X1 X2 X3 X4 X5 X6 X7
S1     NA NA NA NA NA NA NA
S2     NA NA NA NA NA NA NA
S1_S2  NA NA NA NA NA NA NA
S2*    NA NA NA NA NA NA NA
S1_S2* NA NA NA NA NA NA NA

Assuming for run i in the for-loop, column i of the data frame becomes updated:

library(dplyr)

p.list <- vector("list", ncol(accatable))

for(i in seq_along(accatable)){

  accatable[, i] <- testdf[, i] # replace with your actual updating code

  p <- ggplot(accatable[, seq(1, i), drop = FALSE] %>%      # keep only first 1-i columns
           tibble::rownames_to_column(var = "Scenario") %>% # add row name as a column
           tidyr::gather(iteration, value, -Scenario),      # convert to long format
         aes(x = iteration, y = value, group = Scenario,
             color = Scenario, linetype = Scenario)) +
    geom_line() +
    geom_point() +
    labs(x = "Images", y = "ACCA", title = paste("Iteration:", i)) +
    theme_minimal()

  print(p)                       # if you just want to SEE the result from each iteration
  p.list[[i]] <- ggplotGrob(p)   # if you want to SAVE the result from each iteration
}

gridExtra::grid.arrange(grobs = p.list, ncol = 1)

The result looks something like this:

plot

(I've omitted the scale_XX() specifications from the sample code, as I don't think they are essential to the solution. You can adjust the look & feel as needed.)

Upvotes: 1

Related Questions