R, rate of growth, different time - r

Question

I have the following example of dataset for a biology project.
I want to compute a rate of growth of number between 4th January and 2nd of January.
That is rate = (number_at_0104 - number_at_0102)/(number_at_0102)
(In tidyverse if possible)
a <- c("Date", "Specie", "Number")
b <- c("2020-01-01", "Dog", "3")
c <- c("2020-01-02", "Dog", "4")
d <- c("2020-01-03", "Dog", "5")
e <- c("2020-01-04", "Dog", "6")
f <- c("2020-01-01", "Cat", "3")
g <- c("2020-01-02", "Cat", "7")
h <- c("2020-01-03", "Cat", "8")
i <- c("2020-01-04", "Cat", "10")
df <- as.data.frame(rbind(b, c, d, e, f, g, h, i))
names(df) <- a
df$Date <- as.Date(df$Date)
df$Number <- as.integer(df$Number)
I want to calculate a rate of growth. I know this has been treated already but I'm not sure whether I can apply it there.
Usually, we use the lag() function but I have some questions.
Can we tell the lag function what lags to use (eg not the previous periods but 4 periods before)
My dataset is much bigger and for some specie (say cat) I want to compute the rate of growth between 20 Feb and 3 March. And for others (say dog) I want to compute between 5 May and 4th April. How can I do it?
Thank you in advance,

Answer 1

using dplyr
start <- as.Date("2020-01-02")
end <- as.Date("2020-01-04")
df %>%
filter(Date %between% c(start, end)) %>%
arrange(Date, Species) %>%
group_by(Species) %>%
summarise(Growth = (last(Number) - first(Number)) / first(Number))
output
Species Growth
<fct> <dbl>
1 Cat 0.25
2 Dog 0.5
data
note that my test data is already as dates and as numeric values
df <- data.frame(
Date = rep(seq.Date(as.Date("2020-01-01"), as.Date("2020-01-04"), "days"), 2),
Species = c(rep("Dog", 4), rep("Cat", 4)),
Number = 3:10
)
if you want different lookups for every species you can do something like this. Define your lookups and the output will be a tibble with Species, Growth and the period it was taken from.
lookups <- list(
c("Species" = "Dog", "start" = "2020-01-01", "end" = "2020-01-04"),
c("Species" = "Cat", "start" = "2020-01-02", "end" = "2020-01-04")
)
bind_rows(lapply(lookups, function(species) {
df %>%
filter(Species == species["Species"] & Date %between% as.Date(c(species["start"], species["end"]))) %>%
arrange(Date, Species) %>%
group_by(Species) %>%
summarise(
Growth = (last(Number) - first(Number)) / first(Number),
Start = species["start"],
End = species["end"]
)
}))
# # A tibble: 2 x 4
# Species Growth Start End
# <chr> <dbl> <chr> <chr>
# 1 Dog 1 2020-01-01 2020-01-04
# 2 Cat 0.25 2020-01-02 2020-01-04

Answer 2

Since you want to have different input for different species and different start and end dates, I have created a function to do the job.
Function compute_rate_in_species()
x is a dataframe
species is Species in x, which can be supplied in a vector (e.g. c("Dog", "Cat"))
start_date and end_date is self explanatory
DF is a logical value indicating whether you want dataframe as output. If FALSE, a named vector with be the output
If you would like to have different start_date and end_date for different species, you'll have to run the function separately using your desired input.
library(tidyverse)
compute_rate_in_species <- function(x, species, start_date, end_date, DF = T) {
x <- x %>% filter(Species %in% species & Date %in% as.Date(c(start_date, end_date))) %>%
group_by(Species) %>%
summarize(Rate = (last(Number) - first(Number))/first(Number))
if (DF == T) {
return(x)
} else {
x_vec <- setNames(x$Rate, x$Species)
return(x_vec)
}
}
Output
# DF = F
compute_rate_in_species(df, c("Dog", "Cat"), "2020-01-01", "2020-01-03", DF = F)
Cat Dog
1.6666667 0.6666667
# DF = T
compute_rate_in_species(df, c("Dog", "Cat"), "2020-01-01", "2020-01-03", DF = T)
# A tibble: 2 x 2
Species Rate
<chr> <dbl>
1 Cat 1.67
2 Dog 0.667

Answer 3

You can use -
library(dplyr)
start_date <- as.Date("2020-01-02")
end_date <- as.Date("2020-01-04")
df %>%
group_by(Specie) %>%
summarise(growth_rate = (Number[match(end_date, Date)] -
Number[match(start_date, Date)])/
Number[match(start_date, Date)])
# Specie growth_rate
# <chr> <dbl>
#1 Cat 0.429
#2 Dog 0.5
You can replace the start_date and end_date with the dates of your choice.
Or maybe a bit verbose but clearer answer would be -
df %>%
group_by(Specie) %>%
summarise(num_end = Number[match(end_date, Date)],
num_start = Number[match(start_date, Date)],
growth_rate = (num_end - num_start)/num_start)
# Specie num_end num_start growth_rate
# <chr> <int> <int> <dbl>
#1 Cat 10 7 0.429
#2 Dog 6 4 0.5