Looping ggplot across many variables - r

Question

############
EDIT
############
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 you suggested
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?
Lastly, to simplify the y axis I'd like to divide is 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!!!!

Answer 1

You can convert you variable date from character to Date format:
date <- as.Date(date, format = "%Y-%d-%m")
ggplot can print dates at X axes. Now you don't need to create variable months by hand.
I think you should work with data.frame:
df_info <- data.frame(
date = date,
a1 = c(0, 0, 0, 0, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
b1 = c(1, 1, 1, 1, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
hb1 = c(2, 2, 2, 2, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
a2 = c(0, 0, 0, 0, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
b2 = c(1, 1, 1, 1, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
hb2 = c(2, 2, 2, 2, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
a3 = c(0, 0, 0, 0, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
b3 = c(1, 1, 1, 1, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
hb3 = c(2, 2, 2, 2, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
a4 = c(0, 0, 0, 0, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
b4 = c(1, 1, 1, 1, 6421, 41, 5667, 44, 1178, 0, 1070, 1),
hb4 = c(2, 2, 2, 2, 6421, 41, 5667, 44, 1178, 0, 1070, 1)
)
Or if you have big data you can convert matrix to data.frame with converting variables to numeric format (but without names).
df_info <- bind_cols(data.frame(date = date), info_mat[,-1] %>% as.data.frame() %>% lapply(as.numeric)) %>% as.data.frame()
Now we can select columns for first individual:
df <- df_info %>% select(c("date", paste0(c("a", "b", "hb"), 1)))
Next we will create data.frame for plot:
df_plot <- melt(df, id.vars = "date", variable.name = "Type", value.name = "y")
Your function for plot is good you can use it for df_plot. Now let's create function for plotting data for fixed number of individual:
f <- function(num, df_info) {
df <- df_info %>% select(c("date", paste0(c("a", "b", "hb"), num)))
df_plot <- melt(df, id.vars = "date", variable.name = "Type", value.name = "y")
p <- ggplot(df_plot, aes(x = date, y = y, color = `Type`)) + geom_point() +
geom_line() +
labs(
title = "Evolution of a and b and c per months",
subtitle = paste0("plot ", num),
color="Type",
x = "Months",
y = "over months"
)
plot(p)
}
Let's apply our function for each number of individual:
sapply(1:4, f, df_info)
Or
sapply(1:4, function(x) f(x, df_info))
But your data has bad scale. You cannot see the difference between 0 and 1 if you have 6421 on the same plot. but I don't know what you want to do with this data and plots.