论文
image.png
今天的推文我们重复一下论文中的Figure1A
image.png
首先是读取数据
代码语言：javascript复制library(readr)
data.use <- read_csv("use-resistance-seasonality/raw_data/antibiotic_use_data.csv")
head(data.use,n=10)
image.png
对数据进行预处理
代码语言：javascript复制data.use %>%
mutate(year_month =
paste(as.character(year),
as.character(str_pad(month, 2, pad = "0")), sep = "-")) %>%
mutate(x = dense_rank(year_month)) -> f1a_data
这里学到了两个新函数
str_pad可以给字符串补长度，比如
str_pad("ABC",5,pad="0","right")
就可以把ABC补充到5个，结尾补充两个0，这个很有用。比如原来的数字是1,2,3,4,5 改成01,02,03,04,05这种形式
- dense_rank()是做排序的
准备配色 和 添加文本标签的数据
代码语言：javascript复制labels <- data.frame(drug_class = c("Penicillins", "Macrolides", "Quinolones", "Tetracyclines", "Nitrofurans"),
x.pos = c(4, 4, 4, 4, 4),
y.pos = c(6.6, 3.3, 2.4, 1.5, 0.7))
colors <- setNames( c("#220050", "#b30059","#0091a8","#359023", "#ffa500"),
c("Macrolides", "Nitrofurans", "Penicillins", "Quinolones", "Tetracyclines") )
作图
代码语言：javascript复制f1a_data %>%
ggplot(aes(x=x, y=mean_daily_claims_per_10000ppl,
group=drug_class, color=drug_class))
geom_point(size = 0.7)
geom_smooth(aes(fill=drug_class),
span = 0.2, size = 0.7,
method = "loess",
level=0.95,
formula = 'y~x') -> f1a1
print(f1a1)
image.png
下面是细节的美化
代码语言：javascript复制f1a1
geom_text(data = labels, aes(x=x.pos, y=y.pos, label=drug_class), hjust = 0, size = 4.2)
ylab("Mean daily claims/10,000 people")
scale_color_manual(values = colors)
scale_fill_manual(values = colors)
scale_x_continuous(breaks = c(1, 13, 25, 37, 49), labels = c("Jan '11", "Jan '12", "Jan '13", "Jan '14", "Jan '15"))
theme_minimal()
theme(panel.grid.minor = element_blank(),
legend.position = "none",
axis.text = element_text(size = 10),
axis.title.x = element_blank(),
axis.title.y = element_text(size = 11))
image.png
