使用 dplyr 在 R 中进行动态 CAGR 计算

Question

我有以下数据：

 Company    Year    Variables    Data
  ABC        2000     Revenue     10
  ABC        2001     Revenue     15
  ABC        2002     Revenue     12
  ABC        2003     Revenue     25
  ABC        2004     Revenue     30
  CDE        2000     Revenue     5
  CDE        2001     Revenue     8
  CDE        2002     Revenue     17
  CDE        2003     Revenue     9
  CDE        2004     Revenue     34

  #etc

我想计算过去 3 年的复合年增长率 (CAGR)。

例如，每家公司的 3 年复合年增长率结果将是：

Company    Year    Variables    Data    CAGR
 ABC        2000     Revenue     10      NA
 ABC        2001     Revenue     15      NA
 ABC        2002     Revenue     12      6.27%
 ABC        2003     Revenue     25      18.56%
 ABC        2004     Revenue     30     35.72%
 CDE        2000     Revenue     5       NA
 CDE        2001     Revenue     8       NA
 CDE        2002     Revenue     17      50.37%
 CDE        2003     Revenue     9       4.00%
 CDE        2004     Revenue     34      25.99%

我按年份在数据中使用以下公式：

CAGR for 2004=((LastYear/PreviousYear)^(1/n))-1
For example for n = 2
LastYear =2004
PreviousYear =2004-2 = 2002

尝试使用 R 代码计算 2004 年与 2002 年的复合年增长率：

library(tibble)
library(dplyr)
library(lubridate)

year<-c(rep(2000:2004,2))
company<-rep(c("ABC","CDE"),5)
variable<-rep("revenue",10)
data<-c(10,15,12,25,30,5,8,17,9,34)

tibdf<-tibble(company,year,variable,data)
View(tibdf)

#revenue2004<-tibdf%>%filter(year==2004)%>%select(company,data)
#revenue2002<-tibdf%>%filter(year==2001)%>%select(company,data)

计算 CAGR（来自Plot Compound Annual Growth Rate (3 independent variables) in R）

annual.growth.rate <- function(a){

 T1 <- max(a$year) - min(a$year)+1
 FV <- a[which(a$year == max(a$year)),"data"]
 SV <- a[which(a$year == min(a$year)),"data"]
 cagr <- ((FV/SV)^(1/T1)) -1

 }

将 tibdf 用于 in 函数。 不幸的是，我无法将函数应用于我的数据。

感谢您的帮助。

Answer 1

这是一种方法：

library(tidyverse)
df %>%
  arrange(Company, Year) %>%  #in case the years are not in order (here they are)
  group_by(Company) %>%
  mutate(lagY = lag(Year), #get the lag year
         lagD = lag(Data), #get lad Data
         t = Year - lagY, #calculate time
         CAGR = (Data / lagD)^(1/t) - 1) %>% #calculate CAGR
  select(-lagY, -lagD, -t) #remove unwanted variables


#output:
      Company  Year Variables  Data    CAGR
   <fct>   <int> <fct>     <int>   <dbl>
 1 ABC      2000 Revenue      10  NA    
 2 ABC      2001 Revenue      15   0.500
 3 ABC      2002 Revenue      12 - 0.200
 4 ABC      2003 Revenue      25   1.08 
 5 ABC      2004 Revenue      30   0.200
 6 CDE      2000 Revenue       5  NA    
 7 CDE      2001 Revenue       8   0.600
 8 CDE      2002 Revenue      17   1.12 
 9 CDE      2003 Revenue       9 - 0.471
10 CDE      2004 Revenue      34   2.78 

或者在不做中间变量的情况下稍微密集一点：

   df %>%
      arrange(Company, Year) %>%
      group_by(Company) %>%
      mutate(CAGR = (Data/lag(Data))^(1/(Year-lag(Year))) - 1)

数据：

df <- read.table(text ="Company    Year    Variables    Data
ABC        2000     Revenue     10
ABC        2001     Revenue     15
ABC        2002     Revenue     12
ABC        2003     Revenue     25
ABC        2004     Revenue     30
CDE        2000     Revenue     5
CDE        2001     Revenue     8
CDE        2002     Revenue     17
CDE        2003     Revenue     9
CDE        2004     Revenue     34", header = T)

Answer 2

这个函数计算n的不同值的CAGR：

calc_cagr <- function(df, n) {
  df <- df %>%
    arrange(company, year) %>%
    group_by(company) %>%
    mutate(cagr = ((data / lag(data, n)) ^ (1 / n)) - 1)

  return(df)
}

calc_cagr(tibdf, 2)

# A tibble: 10 x 5
# Groups:   company [2]
#    company  year variable  data    cagr
#    <chr>   <int> <chr>    <dbl>   <dbl>
#  1 ABC      2000 revenue  10.0  NA     
#  2 ABC      2001 revenue  15.0  NA     
#  3 ABC      2002 revenue  12.0   0.0954
#  4 ABC      2003 revenue  25.0   0.291 
#  5 ABC      2004 revenue  30.0   0.581 
#  6 CDE      2000 revenue   5.00 NA     
#  7 CDE      2001 revenue   8.00 NA     
#  8 CDE      2002 revenue  17.0   0.844 
#  9 CDE      2003 revenue   9.00  0.0607
# 10 CDE      2004 revenue  34.0   0.414 

然而，我得到的结果与你不同，但你的问题对于是除以 n 还是 n+1 有点模棱两可。

数据

tibdf <- tibble(company = rep(c("ABC", "CDE"), each = 5),
                year = rep(2000:2004, 2),
                variable = rep("revenue", 10),
                data = c(10, 15, 12, 25, 30, 5, 8, 17, 9, 34))