I used this info_mat to compute evolution rates.

date1 <- rbind("February", "March",   "April",     "May",       "June",      "July",      "August",    "September", "October",   "November")
sum1.visit_bush. <- rbind("0", "0" ,"1"  , "-0.75" ,"2","0" ,"0.333333333333333" , "1.25"  , "0",  "-1")
sum1.counts_bush. <- rbind("0" ,"0.115290451813933", "-0.557273997206146", "0.146270002253775" ,  "0.100865119937082", "0.512412930880514",  "0.435049598488427",   "-0.0831961816984858", "0.824791311372408",  "-0.156025577963601" )
sum1.hcounts_bush. <- rbind("0",  "0.0387010676156584", "-0.625695931477516", "0.47254004576659",  "-0.233100233100233", "0.99290780141844" ,  "-0.032536858159634" , "0.349973725696269" , "0.660957571039315",  "-0.341223341926412")
evolution1 <- data.frame(date1, sum1.visit_bush., sum1.counts_bush., sum1.hcounts_bush.)

I then proceed as follows:

df_month_cand <- evolution1 %>% select(c("date", paste0(c("sum.visit_", "sum.counts_", "sum.hcounts_"), "bush.")))
df_month_cand_plot <- melt(df_month_cand, id.vars = "date", variable.name = "Type", value.name = "y")

FunctionPlot <- function(cand, evolution) {
  df_month_cand <- evolution %>% select(c("date1", paste0(c("sum1.visit_", "sum1.counts_", "sum1.hcounts_"), cand)))
  df_month_cand_plot <- melt(df_month_cand, id.vars = "date1", variable.name = "Type", value.name = "y")
  
  p <- ggplot(df_month_cand_plot, aes(x = date1, y = y, color = Type))   geom_point()   geom_line(aes(group=Type))  
    labs(
      title = paste0("Evolution of visits and coverage
      per month for ", cand) ,
      subtitle = "We read: from March to April, whereas the visits of -candidate- increased by -value*100 %-, 
    the coverage in newspapers decreased by -value*100 %-",
      color="Type", 
      x="Months", 
      y="Percentage change over months")   
    theme(
      plot.title = element_text(size=15, face="bold", margin = margin(5, 0, 10, 10), vjust=2, hjust=0.5),
      axis.text.x=element_text(angle=50, size=11.5, vjust=0.5),
      axis.title.y = element_text(vjust=4),
      plot.margin = unit(c(1, 0.3, 0.5, 0.6), "cm"),
      legend.position = "bottom", 
      legend.box.background = element_rect(color="black", size=2), 
      legend.title = element_text(face = "bold", size=10), 
      legend.background = element_rect(fill="grey90",
                                       size=0.5, linetype="solid", 
                                       colour ="black"), 
      panel.background = element_rect(fill = "gray90", colour = "gray70", size = 0.5, linetype = "solid"),
      panel.grid.major = element_line(size = 0.5, linetype = 'dashed', colour = "gray75"))  
    scale_color_manual(labels = c("Visits", "Main text count", "Headline count"), values = c("tomato3", "deepskyblue3", "green2"))   
    scale_x_discrete(limits = c("February", "March", "April", "May", "June", "July", "August", "September", "October", "November"))   
    scale_y_discrete()
  plot(p)
}

sapply("bush.", FunctionPlot, evolution1)

However, on the output the y axis is completely messed-up. The values are not sorted from least to greatest. Why? How to resolve this?

Then, I want to simplify the y axis: I'd like to divide it from -1 to 2 with breaks of 0.25 I tried

scale_y_continuous(breaks=seq(-1, 2, 0.25))

But I have the following error code: Error: Discrete value supplied to continuous scale

Thanks!!!!

CodePudding user response：

None of your data is numeric, all are factors. So you plot against a discrete factor value y and not as desired to numeric continuous scale.

str(evolution1)
# 'data.frame': 10 obs. of  4 variables:
#  $ date1             : Factor w/ 10 levels "April","August",..: 3 6 1 7 5 4 2 10 9 8
#  $ sum1.visit_bush.  : Factor w/ 7 levels "-0.75","-1","0",..: 3 3 5 1 7 3 4 6 3 2
#  $ sum1.counts_bush. : Factor w/ 10 levels "-0.0831961816984858",..: 4 6 3 7 5 9 8 1 10 2
#  $ sum1.hcounts_bush.: Factor w/ 10 levels "-0.032536858159634",..: 5 6 4 8 2 10 1 7 9 3