从 R 中同一图表上的数据集创建多个箱线图

Question

The table below is a small part of a data set I want to use to make three boxplots of:

##         No year month day hour PM2.5  PM10   SO2   NO2   CO    O3  TEMP   PRES
## 1     7345 2014     1   1    0  20.0  90.0  18.0  62.0   NA    NA  -1.5 1007.3
## 2     7346 2014     1   1    1  43.0 348.0  25.0  91.0 1100   1.0  -2.6 1006.9
## 3     7347 2014     1   1    2  79.0 423.0  41.0 103.0 1800   1.0  -3.0 1006.9
## 4     7348 2014     1   1    3  82.0 337.0  43.0 101.0 2100   1.0  -3.3 1006.4
## 5     7349 2014     1   1    4 124.0 594.0  59.0 130.0 2400   1.0  -2.7 1006.1
## 6     7350 2014     1   1    5  89.0 307.0  47.0 102.0 2500   1.0  -3.1 1006.6
## 7     7351 2014     1   1    6  59.0 161.0  45.0  91.0 1900   1.0  -2.6 1007.2
## 8     7352 2014     1   1    7  31.0  93.0  24.0  69.0  900   4.0  -2.9 1007.9
## 9     7353 2014     1   1    8  21.0  90.0  18.0  52.0  700  17.0   0.0 1008.8
## 10    7354 2014     1   1    9  38.0 142.0  25.0  87.0 1200   1.0   9.0 1009.6

I want to make three boxplots of the variable PM2.5 where all of them have the condition hour == 12 but one for each of the years 2014, 2015 and 2016. They also need to be on the same graph. This is what I've tried:

plots <- data.frame(box14 = subset(dat, hour == 12 & year == 2014)$PM2.5,
                    box15 = subset(dat, hour == 12 & year == 2015)$PM2.5,
                    box16 = subset(dat, hour == 12 & year == 2016)$PM2.5)
boxplot(plots)

I just get 'error in data.frame... arguments implying a differing number of rows'. What should I change to get the boxplots I want?

Answer 1

Here's a pretty quick solution using dplyr and ggplot2 :

library(dplyr)
library(ggplot2)

dat %>% 
  as_tibble() %>% 
  filter(hour == 12) %>% 
  filter(year %in% c(2014, 2015, 2016)) %>% 
  ggplot(aes(x=as.factor(year), y=PM2.5)) + 
  geom_boxplot()

Answer 2

In base r:

filtered_data <- original_data[original_data$hour == 12 & original_data$year %in% c(2014,2015,2016),]

and then

boxplot(PM2.5~year,data=filtered_data)