Plotting a graph with multiple geom_lines with loop

Question

I have a dataframe with different percent values of prices variation for a bunch of products over some months. The dataframe is something like this:

 DATA          P10          P25          P50          P75          P90
1  2011-03-01  0.034638180  0.086482130  0.133986300  0.177072700  0.233044900
2  2011-04-01 -0.185378000 -0.112070500 -0.064632480 -0.027086950  0.036643230
3  2011-05-01  0.008258164  0.053702510  0.094340370  0.137678700  0.270847900
4  2011-06-01 -0.105608500 -0.072065040 -0.019818160  0.018149950  0.069389460
5  2011-07-01 -0.080303930 -0.040885830 -0.006315288  0.030778970  0.084747610
6  2011-08-01  0.001524279  0.052229100  0.075928880  0.126691500  0.167735600
7  2011-09-01 -0.097216090 -0.066777680 -0.040682890 -0.014226140  0.034411750

And the code i wrote to create my plot is:

 ggplot()+
      geom_line(data = dataPerc, aes(x = dataPerc$DATA, y =dataPerc$P10,color="P10"),size=1)+
      geom_line(data = dataPerc, aes(x = dataPerc$DATA, y =dataPerc$P25,color='P25'),size=1)+
      geom_line(data = dataPerc, aes(x = dataPerc$DATA, y =dataPerc$P50,color = "P50"),size = 1)+
      geom_line(data = dataPerc, aes(x = dataPerc$DATA, y =dataPerc$P75,color= "P75"),size=1)+
      geom_line(data = dataPerc, aes(x = dataPerc$DATA, y =dataPerc$P90,color="P90"),size=1)+
      scale_x_date(date_labels="%b %y",date_breaks  ="1 month")+
      theme(axis.text.x = element_text(angle = 90))+
      labs(color='Percentile')+
      scale_y_continuous(labels = function(x) paste0(x*100, "%"))+
      xlab("Moth/Year")+
      ylab("% fat. ")

Basically, I want to create the same plot with a loop that substitute the sequence of geom_lines above. Thanks.

Answer 1

Once your data is in wide format, the possibilities are endless. I simplified your code a bit and faceted the data by your original column names (e.g., p10, p25, p50, etc.). This allows you to plot a separate line within each category—independently. Now you can observe the trend from March to September within each facet. I organized the facets into one column. Only the individual month names span the x-axis since you only have data for one year. Feel free to adjust the col = ... argument to find the proper presentation.

If faceting is not your style, then drop the call to facet_wrap() altogether and try inserting col = factor(perc) inside of aes(). This will stack the lines on top of each other on one plot; you also get a nice legend for free. I will demonstrate both methods below.

# Here is how to avoid looping and layering on multiple geoms

library(tidyverse)
library(lubridate)

df <- tribble(
  ~date, ~p10, ~p25,  ~p50, ~p75, ~p90,
  "2011-03-01", 0.034638180, 0.086482130, 0.133986300, 0.177072700, 0.233044900,
  "2011-04-01", -0.185378000, -0.112070500, -0.064632480, -0.027086950, 0.036643230,
  "2011-05-01", 0.008258164, 0.053702510, 0.094340370, 0.137678700, 0.270847900,
  "2011-06-01", -0.105608500, -0.072065040, -0.019818160, 0.018149950, 0.069389460,
  "2011-07-01", -0.080303930, -0.040885830, -0.006315288, 0.030778970, 0.084747610,
  "2011-08-01", 0.001524279, 0.052229100, 0.075928880, 0.126691500, 0.167735600,
  "2011-09-01", -0.097216090, -0.066777680, -0.040682890, -0.014226140, 0.034411750)

# Some quick data preparation

long_df <- df %>%
  mutate(date = ymd(date)) %>%
  pivot_longer(-date, names_to = "perc", values_to = "p_scores")

# Here is a subset of the data frame in long format

# A tibble: 35 x 3
   date       perc  p_scores
   <date>     <chr>    <dbl>
 1 2011-03-01 p10     0.0346
 2 2011-03-01 p25     0.0865
 3 2011-03-01 p50     0.134 
 4 2011-03-01 p75     0.177 
 5 2011-03-01 p90     0.233 
 6 2011-04-01 p10    -0.185 
 7 2011-04-01 p25    -0.112 
 8 2011-04-01 p50    -0.0646
 9 2011-04-01 p75    -0.0271
10 2011-04-01 p90     0.0366
# … with 25 more rows

# Simplified code

ggplot(long_df, aes(x = date, y = p_scores)) +
  geom_line(size = 1) +
  scale_x_date("Month", 
               date_breaks = "1 month", 
               date_labels = '%B') +
  scale_y_continuous("% Fat.", labels = function(x) paste0(x*100, "%")) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
  facet_wrap(~ perc, ncol = 1)

Below is my other recommendation if you want to stack the lines onto one plot. It appears each line moves in tandem over time without much volatility. I included the code as well for reproducibility.

ggplot(long_df, aes(x = date, y = p_scores, col = factor(perc))) +
  geom_line(size = 1) +
  scale_x_date("Month", 
               date_breaks = "1 month", 
               date_labels = '%B') +
  scale_y_continuous("% Fat.", labels = function(x) paste0(x*100, "%")) +
  labs(color = "Score \nType:") +  # This is a generic legend title
  theme_bw() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

Also, you can achieve the same result without converting perc to a factor variable inside of aes(). But I digress.

I hope this helps!

Answer 2

Here's an answer without the loop you said you wanted. ggplot2 does not work well with loops.

# Read in your data - changed `DATA` to `date`
a <- 
"date          P10          P25          P50          P75          P90
1  2011-03-01  0.034638180  0.086482130  0.133986300  0.177072700  0.233044900
2  2011-04-01 -0.185378000 -0.112070500 -0.064632480 -0.027086950  0.036643230
3  2011-05-01  0.008258164  0.053702510  0.094340370  0.137678700  0.270847900
4  2011-06-01 -0.105608500 -0.072065040 -0.019818160  0.018149950  0.069389460
5  2011-07-01 -0.080303930 -0.040885830 -0.006315288  0.030778970  0.084747610
6  2011-08-01  0.001524279  0.052229100  0.075928880  0.126691500  0.167735600
7  2011-09-01 -0.097216090 -0.066777680 -0.040682890 -0.014226140  0.034411750
"
df <- read.table(text = a, header = TRUE)
library(tidyr)
library(dplyr)
library(ggplot2)
# make the data tidy. ggplot2 needs tidy data (one observation per row)
df <- df %>% pivot_longer(cols = -date, names_to = "pct")
# format date as date
df$date <- as.Date(df$date)

ggplot(df, aes(x = date, y = value, color = pct)) +
  geom_line(size=1) +
    scale_x_date(date_labels="%b %y",date_breaks  ="1 month") +
  theme(axis.text.x = element_text(angle = 90)) +
  labs(color='Percentile') +
  scale_y_continuous(labels = function(x) paste0(x * 100, "%")) +
  xlab("Month/Year") +
  ylab("% fat. ")
```[![enter image description here][1]][1]


  [1]: https://i.sstatic.net/OCk8O.png

Answer 3

Don't use a loop - convert your data from wide to long.

long_data = tidyr::pivot_longer(your_data, -DATA, names_to = "Percentile")

ggplot(long_data, aes(x = DATA, y = value, color = name)) +
  geom_line(size = 1) + 
  theme(axis.text.x = element_text(angle = 90)) +
  labs(x = "Month/Year", y = "% fat. ") +
  scale_y_continuous(labels = scales::label_percent(accuracy = 0.1))      

Also, don't use data$column inside aes() - it expects unquoted column names.