conditional average using dplyr in R-CodePudding

I'm working with a cross-country panel dataset.

I want to create a new variable (let's call it C1) such that the variable for Argentina in 1999 reflects the average value of inf from the beginning of the sample till the previous year (i.e., 1998).

Likewise, C1 for Turkey in 1991 reflects the average of inf from the beginning of the sample till 1990. Note that the first year for which data is available might be different for each country.

I would like to ideally do this using dplyr. Any suggestions or leads would be much appreciated.

Below is a snapshot of my data.

# A tibble: 82 x 3
   country    year   inf
   <chr>     <dbl> <dbl>
 1 Argentina  1960 NA   
 2 Argentina  1961 20.3 
 3 Argentina  1962 28.9 
 4 Argentina  1963 25.6 
 5 Argentina  1964 28.8 
 6 Argentina  1965 21.2 
 7 Argentina  1966 25.6 
 8 Argentina  1967 29.0 
 9 Argentina  1968 10.3 
10 Argentina  1969  7.80
# ... with 72 more rows

structure(list(country = c("Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Argentina", "Argentina", 
"Argentina", "Argentina", "Argentina", "Turkey", "Turkey", "Turkey", 
"Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", 
"Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", 
"Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", 
"Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", 
"Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", "Turkey", 
"Turkey", "Turkey", "Turkey"), year = c(1960, 1961, 1962, 1963, 
1964, 1965, 1966, 1967, 1968, 1969, 1970, 1971, 1972, 1973, 1974, 
1975, 1976, 1977, 1978, 1979, 1980, 1981, 1982, 1983, 1984, 1985, 
1986, 1987, 1988, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 
1997, 1998, 1999, 2000, 1960, 1961, 1962, 1963, 1964, 1965, 1966, 
1967, 1968, 1969, 1970, 1971, 1972, 1973, 1974, 1975, 1976, 1977, 
1978, 1979, 1980, 1981, 1982, 1983, 1984, 1985, 1986, 1987, 1988, 
1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 
2000), inf = c(NA, 20.3106975252928, 28.8718416517331, 25.5911543481732, 
28.7746173195633, 21.2329349235212, 25.6034624263967, 29.0182629356081, 
10.2753105574405, 7.79583003660558, 6.46787463937974, 31.2710638825753, 
64.2447859181755, 65.5352873661713, 30.6304413715773, 197.697364540376, 
438.322779724218, 159.427177705708, 161.372172005774, 147.377044038694, 
95.7904246111226, 105.27636637456, 194.535299671424, 380.158528520508, 
611.196303889043, 607.447497820506, 77.2922366864323, 127.539918062845, 
381.246344411434, 3046.09115199662, 2078.31681803114, 140.502378663632, 
16.0719935356945, -3.56109557576575, 2.84933971509302, 3.16512339016516, 
-0.0523754965904146, -0.46391313116429, -1.70527962400085, -1.83655839220664, 
1.03728710073175, NA, 4.37340043959567, 5.78893779037435, 6.4166944802911, 
2.35243651899155, 4.11785566096488, 5.94759067350768, 5.93989052660011, 
4.76015624386707, 6.87631710287759, 8.64450444669336, 16.9020871544943, 
10.9756760490106, 21.9302187016976, 28.9923473546045, 21.3092400682888, 
15.6126574914859, 24.0890232450352, 47.5411467102416, 76.7208667189401, 
93.0032247916414, 44.0570704966979, 28.2268344007077, 26.2582454161426, 
48.2368326072076, 53.054476874505, 36.0068855504432, 33.6122290331458, 
69.0181311101417, 75.4048306122696, 58.2443981066746, 59.1641068909336, 
65.1994382769795, 68.3794283559812, 104.749137215122, 86.007542441817, 
77.223512216706, 81.4548625820616, 143.639661614673, 54.2799583040521, 
49.3607762430324)), row.names = c(NA, -82L), class = c("tbl_df", 
"tbl", "data.frame"))

CodePudding user response：

Use cummean. Since you're first value is NA, you can group_by NA values and country, and apply the cummean.

df %>% 
  group_by(country, na = is.na(inf)) %>% 
  mutate(C1 = head(c(NA, cummean(inf)), -1)) %>% 
  ungroup() %>% 
  select(-na)

# A tibble: 82 x 4
   country    year   inf    C1
   <chr>     <dbl> <dbl> <dbl>
 1 Argentina  1960 NA     NA  
 2 Argentina  1961 20.3   NA  
 3 Argentina  1962 28.9   20.3
 4 Argentina  1963 25.6   24.6
 5 Argentina  1964 28.8   24.9
 6 Argentina  1965 21.2   25.9
 7 Argentina  1966 25.6   25.0
 8 Argentina  1967 29.0   25.1
 9 Argentina  1968 10.3   25.6
10 Argentina  1969  7.80  23.7
# ... with 72 more rows

CodePudding user response：

To deal with the years with missing values, you can compute the running mean in another dataframe with missings dropped, then merge the results back into your original dataframe.

library(dplyr)

avg_calc <- mydata %>% 
  filter(!is.na(inf)) %>% 
  group_by(country) %>% 
  mutate(
    C1 = cummean(inf),
    year = year   1,    # add one to year so merged avg reflects previous year
    inf = NULL
  ) %>% 
  ungroup()

mydata <- mydata %>% 
  left_join(avg_calc)

mydata
#> # A tibble: 82 x 4
#>    country    year   inf    C1
#>    <chr>     <dbl> <dbl> <dbl>
#>  1 Argentina  1960 NA     NA  
#>  2 Argentina  1961 20.3   NA  
#>  3 Argentina  1962 28.9   20.3
#>  4 Argentina  1963 25.6   24.6
#>  5 Argentina  1964 28.8   24.9
#>  6 Argentina  1965 21.2   25.9
#>  7 Argentina  1966 25.6   25.0
#>  8 Argentina  1967 29.0   25.1
#>  9 Argentina  1968 10.3   25.6
#> 10 Argentina  1969  7.80  23.7
#> # ... with 72 more rows