Graphs and Save multiple filter data

I'm working with google analytics data. I have 100 different google analytics campaigns and I have to make a graph for each of them, clearly I can't use facet_grid or wrap due to the number of campaigns.

So I need to make a graph using only the data from one campaign, then save the png of it and make a new graph of another campaign and save the png, and so on.

The code of graphs is:

ggplot(Data%>%filter(Cost < 2000000, Impressions < 400000, Campaign == "CAMPAIGN 1"), 
    aes( x = Cost, color = CTR, y = Conversions, size = Impressions, shape = Week)) 
geom_point() 
scale_size(range = c(0, 10)) 
geom_smooth(se = F, show.legend = F, method = "lm") 
labs(title = "Daily Conversions CAMPAIGN 1",
     shape = "") 
theme_bw() 
scale_color_viridis_b()

I need that in the filter function and in labs change the filter of the campaign and the title, later save the graph in a png with the name of the Campaign.

I know that with a for function it can work but I don't know how to make it do the graph for each of the campaigns.

Thanks in advance

CodePudding user response：

What about this?

 for (i in 1:5) {
name <- paste("Campaign", i)
  
  plot <- ggplot(Data%>%filter(Cost < 2000000, Impressions < 400000, Campaign == name), 
         aes( x = Cost, color = CTR, y = Conversions, size = Impressions, shape = Week)) 
    geom_point() 
    scale_size(range = c(0, 10)) 
    geom_smooth(se = F, show.legend = F, method = "lm") 
    labs(title = "Daily Conversions CAMPAIGN 1",
         shape = "") 
    theme_bw() 
    scale_color_viridis_b()
  
  ggsave(plot, paste0("./", name, ".png"))
}

CodePudding user response：

Consider by, object-oriented wrapper to tapply, that allows you to slice a data frame by factor(s) and run processes on the subsets to return a simplified object (i.e., vector, matrix), or a list of any output:

build_plot <- function(sub) {
    plt <- ggplot(filter(sub, Cost < 2000000, Impressions < 400000), 
              aes(x = Cost, color = CTR, y = Conversions, size = Impressions, shape = Week))  
              geom_point()  
              scale_size(range = c(0, 10))  
              geom_smooth(se = FALSE, show.legend = FALSE, method = "lm")  
              labs(title = paste("Daily Conversions", sub$Campaign[1]), shape = "")  
              theme_bw()  
              scale_color_viridis_b()

    ggsave(plot=plt, filename=paste0(sub$Campaign[1], ".png"), units="in", width=12, height=8)

    return(plt)
}

plot_list <- by(Data, Data$Campaign, build_plot)

CodePudding user response：

At the end, i wrote this code and work fine:

List <- c(unique(Data$`Campaign name`))
List

plot_list = list()

for (i in 1:100) {
  p = ggplot(Data%>%filter(`Campaign name` == List[i]), aes(x=Cost, y=Conversions))   geom_point() 
    labs(title = "Conversiones Diarias según Inversión",
         subtitle = List[i])
  plot_list[[i]] = p
  ggsave(plot_list[[i]], file=paste0("plot_", i,".png"), width = 32, height = 25, units = "cm")
}

Another option without ggsave part in for loop is run this for loop after:

pdf("plots.pdf")
for (i in 1:100) {
  print(plot_list[[i]])
}
dev.off()