Calculations
calculations.Rmdadd_percent()
One of the most used calculations is proportion. If you do not want
to multiply proportion by 100, add argument
multiply = FALSE.
linelist |>
mutate(race_ethnicity = recode_race(Ethnicity, Race, abbr_names = TRUE)) |>
count(race_ethnicity) |>
mutate(percent = add_percent(n, digits = 1))
#> # A tibble: 9 × 3
#> race_ethnicity n percent
#> <chr> <int> <dbl>
#> 1 AI/AN 10 9.5
#> 2 Asian 7 6.7
#> 3 Black/African American 16 15.2
#> 4 Hispanic/Latinx 19 18.1
#> 5 Missing/Unknown 8 7.6
#> 6 Multiple Races 6 5.7
#> 7 NHOPI 10 9.5
#> 8 Other 15 14.3
#> 9 White 14 13.3n_percent()
When preparing data for {ggplot2}, you may want to create labels with
n and percent. Use argument n_suppress to suppress low
values and reverse to set order of n and % (default is n
then %).
linelist |>
mutate(race_ethnicity = recode_race(Ethnicity, Race, abbr_names = TRUE)) |>
count(race_ethnicity) |>
mutate(
percent = add_percent(n, digits = 1),
label = n_percent(n, percent, reverse = TRUE, n_suppress = 10)
)
#> # A tibble: 9 × 4
#> race_ethnicity n percent label
#> <chr> <int> <dbl> <chr>
#> 1 AI/AN 10 9.5 9.5% (10)
#> 2 Asian 7 6.7 **
#> 3 Black/African American 16 15.2 15.2% (16)
#> 4 Hispanic/Latinx 19 18.1 18.1% (19)
#> 5 Missing/Unknown 8 7.6 **
#> 6 Multiple Races 6 5.7 **
#> 7 NHOPI 10 9.5 9.5% (10)
#> 8 Other 15 14.3 14.3% (15)
#> 9 White 14 13.3 13.3% (14)dashboard_tbl()
Pre-aggregating data for dashboarding can be a tedious and
disorganized task. One way to aggregate and organize your data is to
summarize all your data into one table. To use
dashboard_tbl, start with your data, then select which
columns you want to summarize, then call the function:
linelist %>%
mutate(
race_ethnicity = recode_race(Ethnicity, Race),
agegroups = age_groups(Age),
Gender = recode_gender(Gender)
) %>%
select(Gender, agegroups, race_ethnicity) %>%
dashboard_tbl()
#> Variable Category n Percent Label
#> 1 Gender Female 51 48.6 48.6% (51)
#> 2 Gender Male 39 37.1 37.1% (39)
#> 3 Gender Missing/Unknown 15 14.3 14.3% (15)
#> 4 agegroups 0-9 12 11.4 11.4% (12)
#> 5 agegroups 10-19 11 10.5 10.5% (11)
#> 6 agegroups 20-29 11 10.5 10.5% (11)
#> 7 agegroups 30-39 13 12.4 12.4% (13)
#> 8 agegroups 40-49 12 11.4 11.4% (12)
#> 9 agegroups 50-59 12 11.4 11.4% (12)
#> 10 agegroups 60-69 6 5.7 5.7% (6)
#> 11 agegroups 70-79 17 16.2 16.2% (17)
#> 12 agegroups 80+ 11 10.5 10.5% (11)
#> 13 agegroups Missing/Unknown 0 0.0 0% (0)
#> 14 race_ethnicity American Indian/Alaska Native 10 9.5 9.5% (10)
#> 15 race_ethnicity Asian 7 6.7 6.7% (7)
#> 16 race_ethnicity Black/African American 16 15.2 15.2% (16)
#> 17 race_ethnicity Hispanic/Latinx 19 18.1 18.1% (19)
#> 18 race_ethnicity Missing/Unknown 8 7.6 7.6% (8)
#> 19 race_ethnicity Multiple Races 6 5.7 5.7% (6)
#> 20 race_ethnicity Native Hawaiian/Other Pacific Islander 10 9.5 9.5% (10)
#> 21 race_ethnicity Other 15 14.3 14.3% (15)
#> 22 race_ethnicity White 14 13.3 13.3% (14)It is important to remember you only pass the columns you want
summarized into the function, otherwise it will attempt to summarize all
columns. If you’re wanting to summarize by group or year, use
group_by and include that column in the group of columns
you pass into the function. To adjust digits on percent via
digits argument, and change n % order using
reverse.
rate_per_100k()
Another common calculation for our team are is incidence rate per 100,000.
linelist |>
mutate(race_ethnicity = recode_race(Ethnicity, Race, abbr_names = TRUE)) |>
filter(race_ethnicity %in% c("White","NHOPI","Multiple Races")) |>
count(race_ethnicity) |>
mutate(incidence_rate = case_when(
race_ethnicity == "White" ~ rate_per_100k(n, 1500000, digits = 1),
race_ethnicity == "NHOPI" ~ rate_per_100k(n, 100000, digits = 1),
race_ethnicity == "Multiple Races" ~ rate_per_100k(n, 200000, digits = 1)
)
)
#> # A tibble: 3 × 3
#> race_ethnicity n incidence_rate
#> <chr> <int> <dbl>
#> 1 Multiple Races 6 3
#> 2 NHOPI 10 10
#> 3 White 14 0.9time_between()
time_between() calculates the time elapsed between two
dates in the following units: days, weeks, months and years. Value is
always rounded down to account for complete elapsed time. Note: order of
dates does not necessarily matter, but beware of sign.
episode_date = as.Date("2020-04-01")
death_date = as.Date("2020-05-15")
time_between(death_date, episode_date, unit = "days")
#> [1] 44
time_between(episode_date, death_date, unit = "days")
#> [1] -44
dob = as.Date("1986-04-01")
vaccine_date = as.Date("2023-05-15")
time_between(vaccine_date, dob, unit = "years")
#> [1] 37
time_between(dob, vaccine_date, unit = "years")
#> [1] -37