如何使用 R 计算观测值之间的比率

Question

考虑到我有一个按日期排序的数据框，并且对于每个数据框我都有一些数量，我该如何计算每一行的X_天/X_天-1指数？

我的数据集： https : //raw.githubusercontent.com/imdevskp/covid_19_jhu_data_web_scrap_and_cleaning/master/covid_19_clean_complete.csv

我的过程数据集（R 代码）：

library(tidyverse)
library(lubridate)

covid19 <- read.table(file = "covid_19_clean_complete.csv",
                      header = TRUE,
                      stringsAsFactors = FALSE,
                      sep = ",",
                      dec = ".",
                      quote = "\"")

covid19$Date <- mdy(covid19$Date)

brasil <- covid19 %>%
    filter(Country.Region == "Brazil") %>%
    group_by(Country.Region, Date) %>%
    summarise(Cases = sum(Confirmed))

我的费率将根据Cases变量计算。

Answer 1

我们可以采用“案例”的lag并用它来划分“案例”

library(dplyr)
out <- covid19 %>%
          group_by(Country.Region, Date) %>%
          summarise(Cases = sum(Confirmed)) %>%
          mutate(Ratio = Cases/lag(Cases))
out %>%
   filter(Country.Region == "Brazil") %>%
   tail
# A tibble: 6 x 4
# Groups:   Country.Region [1]
#  Country.Region Date       Cases Ratio
#  <chr>          <date>     <int> <dbl>
#1 Brazil         2020-03-08    20  1.54
#2 Brazil         2020-03-09    25  1.25
#3 Brazil         2020-03-10    31  1.24
#4 Brazil         2020-03-11    38  1.23
#5 Brazil         2020-03-12    52  1.37
#6 Brazil         2020-03-13   151  2.90