如何安排从for循环创建的ggplots

Question

I have the following df(sample only)

>ts_data_1
                Month drepo drevrepo dbankrate  dCRR dCallrate  dWPI dGDP  dFED dwidth
2001-05-01 2001-05-01 -0.25    -0.25       0.0 -0.50      0.54  0.19  0.0 -0.50      0
2001-06-01 2001-06-01 -0.25     0.00       0.0  0.00     -0.79 -0.30  0.0 -0.25    -25
2001-07-01 2001-07-01  0.00     0.00       0.0  0.00     -0.05 -0.07  0.7  0.00      0
2001-08-01 2001-08-01  0.00     0.00       0.0  0.00     -0.25  0.18  0.0 -0.25      0
2001-09-01 2001-09-01  0.00     0.00       0.0  0.00      0.36 -0.89  0.0 -0.50      0
2001-10-01 2001-10-01  0.00     0.00      -0.5  0.00      0.10 -1.61  1.5 -0.50      0
2001-11-01 2001-11-01  0.00     0.00       0.0 -1.75     -0.43 -0.32  0.0 -0.50      0
2001-12-01 2001-12-01  0.00     0.00       0.0 -0.25      0.11 -0.51  0.0 -0.25      0
2002-01-01 2002-01-01  0.00     0.00       0.0  0.00     -0.45 -0.57 -0.4  0.00      0
2002-02-01 2002-02-01  0.00     0.00       0.0  0.00      0.10 -0.12  0.0  0.00      0
              dnse        dusd
2001-05-01   42.65  0.13352941
2001-06-01  -60.00  0.08450000
2001-07-01  -35.05  0.13731818
2001-08-01  -19.10 -0.01481818
2001-09-01 -139.90  0.51900000
2001-10-01   58.05  0.37447619
2001-11-01   95.25 -0.02310777
2001-12-01   -8.10 -0.07473684
2002-01-01   16.35  0.39258581
2002-02-01   66.65  0.37628261

I have run a for loop to plot all the time series one by one :

for(i in tail(colnames(ts_data_1), -1)){
gg <- ggplot(ts_data_1, aes_string(x = "Month", y = i)) +
geom_line(size=1.05,colour="#0072B2") + scale_x_date(date_breaks = "2 year",date_labels = "%Y")+ xlab('\nYear') + ggtitle(paste(i,"(Differenced Timeseries)\n")) + theme_bw()
print(gg)
}

This plots graph one below another. I looking for a way to plot all in sort of a grid of 2 column in a single plot frame. Had looked into grid.arrange as well as plot_grid fn of cowplot package but cannot make it work in the loop. Need help here!

PS: Data for reproducibality

ts_data_1<-structure(list(Month = structure(c(11443, 11474, 11504, 11535, 
11566, 11596, 11627, 11657, 11688, 11719), class = "Date"), drepo = c(-0.25, 
-0.25, 0, 0, 0, 0, 0, 0, 0, 0), drevrepo = c(-0.25, 0, 0, 0, 
0, 0, 0, 0, 0, 0), dbankrate = c(0, 0, 0, 0, 0, -0.5, 0, 0, 0, 
0), dCRR = c(-0.5, 0, 0, 0, 0, 0, -1.75, -0.25, 0, 0), dCallrate = c(0.539999999999999, 
-0.789999999999999, -0.0499999999999998, -0.25, 0.359999999999999, 
0.100000000000001, -0.430000000000001, 0.11, -0.45, 0.100000000000001
), dWPI = c(0.19, -0.3, -0.0699999999999994, 0.18, -0.890000000000001, 
-1.61, -0.32, -0.51, -0.57, -0.12), dGDP = c(0, 0, 0.7, 0, 0, 
1.5, 0, 0, -0.399999999999999, 0), dFED = c(-0.5, -0.25, 0, -0.25, 
-0.5, -0.5, -0.5, -0.25, 0, 0), dwidth = c(0L, -25L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L), dnse = c(42.6500000000001, -60, -35.0500000000002, 
-19.0999999999999, -139.9, 58.05, 95.2500000000001, -8.10000000000014, 
16.3500000000001, 66.6499999999999), dusd = c(0.133529411764705, 
0.0844999999999985, 0.137318181818202, -0.0148181818181996, 0.518999999999998, 
0.374476190476202, -0.0231077694236035, -0.0747368421051959, 
0.392585812356998, 0.376282608695597)), .Names = c("Month", "drepo", 
"drevrepo", "dbankrate", "dCRR", "dCallrate", "dWPI", "dGDP", 
"dFED", "dwidth", "dnse", "dusd"), row.names = c("2001-05-01", 
"2001-06-01", "2001-07-01", "2001-08-01", "2001-09-01", "2001-10-01", 
"2001-11-01", "2001-12-01", "2002-01-01", "2002-02-01"), class = "data.frame")

Answer 1

The task becomes easier if you convert your data into long format first, eg, using gather from package tidyr .

You can then do the following.

Plot each column in a separate facet

require(ggplot2)
require(tidyr)

data %>% 
  gather(variable,value,-Month) %>% 
  ggplot(aes(x=Month,y=value)) + facet_wrap(~variable) + geom_line()

If you want that all plots have their own y-scale, then add scales="free_y" to the call to facet_wrap :

ts_data_1 %>% 
  gather(variable, value, -Month) %>% 
  ggplot(aes(x=Month,y=value)) + geom_line() + facet_wrap(~variable,scales="free_y")

Have all columns color-coded within a single plot

data %>%
  gather(variable,value,-Month) %>% 
  ggplot(aes(x=Month,y=value,color=variable)) + geom_line()

Answer 2

While I think ziggystar's answer is usually the correct way to go about plotting in ggplot, I would like to show how to perform the requested operation also.

library(gridExtra)
library(tidyverse)

Create a list of plots and pass that list to arrangeGrob . Since you want the column names as part of the plot titles you can map over them:

Names <- names(ts_data_1)

Names[-1] %>%
  map(function(x){
    ts_data_1 <- data.frame(Month = ts_data_1[,1], y = ts_data_1[[x]])
    pl <- ggplot(ts_data_1, aes(x = Month, y = y)) +
      geom_line(size = 1.05, colour = "#0072B2") + 
      scale_x_date(date_breaks = "2 months", date_labels = "%m") + xlab('\nYear') + 
      ggtitle(paste(x,"\n(Differenced Timeseries)")) + theme_bw()
    return(pl)
  }) %>%
  arrangeGrob(grobs = .) %>% #add ncol/nrow argument here
  grid.arrange()

I changed the x scale to date_breaks = "2 months", date_labels = "%m" .

In order to change the number of cols to 2, add parameter cols = 2 to arrangeGrob .