Expanding Data Frame with cumsum in R - r

Question

I've got a data frame with historc F1 data that looks like this:
Driver
Race number
Position
Number of Career Podiums
Farina
1
1
1
Fagioli
1
2
1
Parnell
1
3
1
Fangio
2
1
1
Ascari
2
2
1
Chiron
2
3
1
...
...
...
...
Moss
47
1
4
Fangio
47
2
23
Kling
47
3
2
now I want to extend it in a way that for every race there is not only the top 3 of that specific Race but also everyone that has had a top 3 before so I can create a racing bar chart. The final data frame should look like this
Driver
Race number
Position
Number of Career Podiums
Farina
1
1
1
Fagioli
1
2
1
Parnell
1
3
1
Fangio
2
1
1
Ascari
2
2
1
Chiron
2
3
1
Farina
2
NA
1
Fagioli
2
NA
1
Parnell
2
NA
1
Parsons
3
1
1
Holland
3
2
1
Rose
3
3
1
Farina
3
NA
1
Fagioli
3
NA
1
Parnell
3
NA
1
Fangio
3
NA
1
Ascari
3
NA
1
Chiron
3
NA
1
Is there any easy way to do this? I couldnt find someone with a similar problem on google.

Answer 1

If I correctly understand your problem, you have only observations for the top3 drivers of every race. But you want to have observations for every driver that has ever achieved a top3 position in your dataset across all races.
For example in the following dataset, driver D only has an observation for the second race where they achieved the first place, but not the other races:
dat <- data.frame(driver = c("A", "B", "C", "D", "A", "B", "B", "A", "C"),
race_number = rep(1:3, each = 3),
position = rep(1:3, 3))
print(dat)
driver race_number position
1 A 1 1
2 B 1 2
3 C 1 3
4 D 2 1
5 A 2 2
6 B 2 3
7 B 3 1
8 A 3 2
9 C 3 3
To add entries for driver D for the races number 1 and 2 you could use tidyr's expand() function or if you want to use base R you could achieve the same using expand.grid() and unique(). This would leave you with a dataframe object containing all possible combinations between the drivers and the race numbers. Afterwards you simply have to left or right join the result with the initial dataframe.
A solution using standard tidyverse packages tidyr and dplyr could look like this:
library(dplyr)
library(tidyr)
dat %>%
expand(driver, race_number) %>%
left_join(dat)
# A tibble: 12 x 4
driver race_number position previous_podium_positions
<chr> <int> <int> <dbl>
1 A 1 1 1
2 A 2 2 2
3 A 3 2 3
4 B 1 2 1
5 B 2 3 2
6 B 3 1 3
7 C 1 3 1
8 C 2 NA NA
9 C 3 3 2
10 D 1 NA NA
11 D 2 1 1
12 D 3 NA NA
Note that the "new" observations will naturally have NAs for the position and the number of previous podium positions. The latter could be added easily via the following approach, which counts the previous
dat %>%
expand(driver, race_number) %>%
left_join(dat) %>%
arrange(race_number) %>%
mutate(previous_podium_positions = ifelse(is.na(previous_podium_positions),0,1)) %>%
group_by(driver) %>%
mutate(previous_podium_positions = cumsum(previous_podium_positions))
Joining, by = c("driver", "race_number")
# A tibble: 12 x 4
# Groups: driver [4]
driver race_number position previous_podium_positions
<chr> <int> <int> <dbl>
1 A 1 1 1
2 B 1 2 1
3 C 1 3 1
4 D 1 NA 0
5 A 2 2 2
6 B 2 3 2
7 C 2 NA 1
8 D 2 1 1
9 A 3 2 3
10 B 3 1 3
11 C 3 3 2
12 D 3 NA 1
I hope this helped. Just a brief disclaimer, these may very well be not the most resource or time-efficient solutions but rather the fastest/easiest way to solve the issue.