BMisc includes miscellaneous functions for working with panel data, quantiles, dealing with formulas, among other things.
You can install BMisc from github with:
# install.packages("devtools")
devtools::install_github("bcallaway11/BMisc")or from CRAN with:
install.packages("BMisc")The make_dist creates a distribution function from a vector of value
of the random variable and the corresponding value of its cdf.
library(BMisc)
y <- rnorm(100)
y <- y[order(y)]
u <- runif(100)
u <- u[order(u)]
F <- make_dist(y, u)
class(F)
#> [1] "ecdf" "stepfun" "function"
# plot(F)Another useful function is the make_balanced_panel function which
drops observations from a panel dataset which are not available in all
time periods.
id <- rep(seq(1, 100, 1), 2) ## individual ids for setting up a two period panel
t <- rep(seq(1, 2), 100) ## time periods
y <- rnorm(200) ## outcomes
dta <- data.frame(id = id, t = t, y = y) ## make into data frame
dta <- dta[-7, ] ## drop the 7th row from the dataset (which creates an unbalanced panel)
nrow(dta)
#> [1] 199
dta <- make_balanced_panel(dta, idname = "id", tname = "t")
nrow(dta) ## now all the observations with missing data in any period are dropped
#> [1] 198