Example of using ggplot2's %+% to follow DRY principle

ggplot2_DRY

library(eurostat)
library(dplyr)
library(ggplot2)
library(scales)

#### Original version (not using %+%) ####
# Originally posted at http://www.r-bloggers.com/european-debt-and-interest/ on June 7, 2015

r1 <- get_eurostat('gov_10dd_edpt1')

# add country names
r2 <- get_eurostat_dic('geo') %>%
    mutate(.,
           geo=V1,
           country=V2,
           country=gsub('\\(.*$','',country)) %>%
    select(.,geo,country) %>%
    merge(.,r1) %>%
    # filter countries
    filter(.,
           !grepl('EA.*',geo),
           !grepl('EU.*',geo),
           geo!='NO')

filter(r2,
       sector=='S13', # general government
       na_item=='B9', # Net lending (+) /net borrowing (-)
       unit=='PC_GDP' # % GDP
) %>%
    ggplot(.,aes(x=time,y=values)) + 
    geom_line()+
    ylab('% GDP')+
    facet_wrap(~country,nrow=4) +
    ggtitle('Net lending (+) /net borrowing (-)') +
    xlab('Year') +
    geom_hline(yintercept=-3,colour='red') +
    scale_x_date(
        breaks=c(as.Date("2000-01-01"),as.Date("2010-01-01") )
        ,labels = date_format("%Y"))

#########
filter(r2,
       sector=='S13', # general government
       na_item=='GD', # Gross debt
       unit=='PC_GDP' # % GDP
) %>%
    ggplot(.,aes(x=time,y=values)) + 
    geom_line()+
    ylab('% GDP')+
    facet_wrap(~country,nrow=4) +
    ggtitle('Gross Debt') +
    xlab('Year') +
    geom_hline(yintercept=60,colour='red') +
    scale_x_date(
        breaks=c(as.Date("2000-01-01"),as.Date("2010-01-01") )
        ,labels = date_format("%Y"))

#########
filter(r2,
       sector=='S13', # general government
       na_item=='D41PAY', # Interest, payable
       unit=='PC_GDP' # % GDP
) %>%
    ggplot(.,aes(x=time,y=values)) +
    geom_line()+
    ylab('% GDP')+
    facet_wrap(~country,nrow=4) +
    ggtitle('Interest, payable') +
    xlab('Year') +
    scale_x_date(
        breaks=c(as.Date("2000-01-01"),as.Date("2010-01-01") )
        ,labels = date_format("%Y"))

#### My version (using %+%) ####

r1 <- get_eurostat('gov_10dd_edpt1')

# add country names
r2 <- get_eurostat_dic("geo") %>%
    mutate(geo     = V1,
           country = V2,
           country = str_replace(country, "\\(.*$","")) %>%
    select(geo, country) %>%
    inner_join(r1) %>%
    filter(!str_detect(geo, "[0-9]$"),
           geo    != "NO",
           sector == "S13",
           unit   == "PC_GDP") %>% 
    tbl_df()

p_base <- r2 %>% 
    ggplot(aes(x = time, y = values)) + 
    geom_line() +
    facet_wrap(~country, nrow = 4) +
    scale_x_date(breaks = as.Date(c("2000-01-01", "2010-01-01")),
                 labels = date_format("%Y")) +
    labs(x = "Year", y = "% GDP")

p1 <- p_base %+% filter(r2, na_item == "B9") + 
    geom_hline(yintercept = -3, colour = "red") +
    ggtitle("Net lending (+) /net borrowing (-)")
    
p2 <- p_base %+% filter(r2, na_item == "GD") + 
    geom_hline(yintercept = 60, colour = "red") +
    ggtitle("Gross Debt")

p3 <- p_base %+% filter(r2, na_item == "D41PAY") +
    ggtitle("Interest, payable")

#### Using function instead of %+% ####

make_plot <- function(df = r2, plot_var, title) {
    df %>% 
        filter(na_item == plot_var) %>% 
        ggplot(aes(x = time, y = values)) + 
        geom_line() +
        facet_wrap(~country, nrow = 4) +
        scale_x_date(breaks = as.Date(c("2000-01-01", "2010-01-01")),
                     labels = date_format("%Y")) +
        labs(x = "Year", y = "% GDP", title = title)
}

p1 <- make_plot(plot_var = "B9", title = "Net lending (+) /net borrowing (-)") +
    geom_hline(yintercept = -3, colour = "red")
    
p2 <- make_plot(plot_var = "GD", title = "Gross Debt") +
    geom_hline(yintercept = 60, colour = "red")

p3 <- make_plot(plot_var = "D41PAY", title = "Interest, payable")

#### Benchmarking ####
library(microbenchmark)

microbenchmark(
    p3 <- p_base %+% filter(r2, na_item == "D41PAY") +
        ggtitle("Interest, payable"),
    p3 <- make_plot(plot_var = "D41PAY", title = "Interest, payable")
    )

# Unit: milliseconds
# expr      min      lq       mean     median   uq       max        neval
# %+%       3.911188 4.076719 4.241338 4.147733 4.215931 12.39172   100
# make_plot 7.997508 8.245144 8.754945 8.444444 8.579021 16.44030   100