R/basepop.R
In timriffe/DemoTools: Standardize, Evaluate, and Adjust Demographic Data
Defines functions
lt_infer_radix_from_1L0
ArgsCheck
basepop_five
Documented in
basepop_five
#' BPA and BPE methods for adjusting age groups under 10
#' @description Adjust population counts for the age groups 0 to 10
#'
#' @details
#'
#' \code{basepop_five} and \code{basepop_single} can estimate both the BPA and
#' BPE methods. If the user specifies \code{SmoothedFemales}, both
#' \code{basepop_*} functions will return the BPA method.
#' If \code{SmoothedFemales} is left empty, both \code{basepop_*} functions will
#' adjust using the BPE method.
#'
#' For \code{basepop_five}, adjusting the female population counts is the
#' default. For this, only the \code{location}, \code{refDate} and
#' \code{Females_five} are needed. All other arguments are downloaded
#' or set to sensible defaults. For adjusting the male population
#' counts, the user needs to specify the \code{Males_five} population
#' counts and set \code{female = FALSE}.
#'
# # For \code{basepop_single}, the same procedure applies. The only difference
# # is that the vector \code{Males_five} is named \code{Males_single} and accepts
# # a vector of single ages rather than five year abridged age groups. Similarly,
# # the vector for females is \code{Females_single} rather than
# # \code{Females_five} and accepts single age groups.
#'
#' Currently, \code{basepop_five} works only with five year abridged age groups
# #' while \code{basepop_single} works only with single year age groups.
#'
#' The BPE method is used by default. To adjust the counts using
#' the BPA method, the user needs to provide the \code{SmoothedFemales}
#' argument. This is the female population counts passed through
#' a smoothing function such as \code{smooth_age_5}. See the examples
#' section for some examples.
#'
#' @section BPA:
#'
#' Description:
#'
#' The method estimates a smoothed population ages 10 and over and adjusts
#' the population under age 10 using the smoothed population and estimates
#' of fertility and mortality.
#'
#' Based on the smoothed female population counts, it rejuvenates the female
#' "reported" population 20 to 59 years of age for the two 5 year periods prior
#' to the census date to represent the female population in reproductive ages
#' 5 and 10 years earlier. Based on the rejuvenated population and fertility
#' and mortality levels, the method then estimates the male and female births
#' during the two 5 year periods prior to the census date. Next, it projects
#' the two 5-year birth cohorts to the census date. The projected figures
#' represent the adjusted population ages 0 to 4 years and 5 to 9 years
#' at the census date.
#'
#' Advantages:
#'
#' (1) The method adjusts under-10 population to be consistent with fertility
#' and mortality levels and adjusted adult female population.
#'
#' Limitations:
#'
#' (1) BPA assumes a linear change in fertility and mortality during the decade
#' prior to the reference year.
#'
#' (2) The procedure ignores migration, which can lead to misleading results.
#' There are two issues. First, age groups 0-4 and 5-9 are subject to migration,
#' which would affect the comparability of estimated and reported populations
#' in the base year. Second, the estimated size of age groups 0-4 and 5-9 are
#' calculated from numbers of women of reproductive age in the base year
#' rejuvenated to points in the past. With migration, rejuvenated number of
#' women may exceed or be smaller than the number present, and giving
#' birth to children, in the decade prior to the base year.
#'
#' (3) BPA’s smoothing calculations may mask unusual, but real, variations
#' in population age group size. Smoothing irregularities in age structure
#' not attributable to age misreporting will distort estimated births and
#' survived children in the base year.
#'
#' Assumptions:
#'
#' (1) No significant international migration took place within the
#' reference periods for the population, mortality, and fertility input.
#'
#' (2) The data input as the "reported" population is not affected by
#' underenumeration of persons in certain ages, nor by age misreporting.
#'
#' @section BPE:
#'
#' Description:
#'
#' The method adjusts the population under age 10 using the reported population
#' ages 10 and above and estimates of fertility and mortality.
#'
#' The method rejuvenates the reported female population 20 to 59 years of age
#' for the two 5 year periods prior to the census date to represent the female
#' population in reproductive ages 5 and 10 years earlier. Based on the
#' rejuvenated population and fertility and mortality levels, the method then
#' estimates the male and female births during the two 5 year periods prior to
#' the census date. Next, it projects the two 5-year birth cohorts to the
#' census date. The projected figures represent the adjusted population ages
#' 0 to 4 years and 5 to 9 years at the census date.
#'
#' Advantages:
#'
#' (1) The method adjusts the under-10 population to be consistent with
#' fertility and mortality levels and adult female population.
#'
#' Limitations:
#'
#' (1) BPE assumes a linear change in fertility and mortality during the decade
#' prior to the reference year.
#'
#' (2) The procedure ignores migration, which can lead to misleading results.
#' There are two issues. First, age groups 0-4 and 5-9 are subject to
#' migration, which would affect the comparability of estimated and reported
#' populations in the base year. Second, the estimated size of age groups
#' 0-4 and 5-9 are calculated from numbers of women of reproductive age in
#' the base year rejuvenated to points in the past. With migration, rejuvenated
#' number of women may exceed or be smaller than the number present, and
#' giving birth to children, in the decade prior to the base year.
#'
#' (3) The method does not adjust for possible underenumeration and age
#' misreporting errors in the over-10 “reported” population. If the
#' reported population is subject to age-misreporting or age-sex-specific
#' underenumeration, the over-10 population should be smoothed or otherwise
#' corrected prior to use.
#'
#' Assumptions:
#'
#' (1) No significant international migration took place within the reference
#' periods for the population, mortality, and fertility input.
#'
#' (2) The data input as the “reported” population is not affected by
#' underenumeration of persons in certain ages, nor by age misreporting.
#'
#' @return `basepop_five` returns a list with the following elements:
#' *
#' * `Females_adjusted` numeric vector of adjusted population counts for females. Age groups 0, 1-4, and 5-9 are adjusted, while ages 10 and higher are unchanged.
#' * `Males_adjusted` numeric vector of adjusted population counts for males. Age groups 0, 1-4, and 5-9 are adjusted, while ages 10 and higher are unchanged.
#' * `Females_five` numeric vector of female population counts given as input.
#' * `Males_five` numeric vector of male population counts given as input.
#' * `nLxf` numeric matrix of female `nLx`, abridged ages in rows and (potentially interpolated) time in columns. Potentially downloaded.
#' * `nLxm` numeric matrix of male `nLx`, abridged ages in rows and (potentially interpolated) time in columns. Potentially downloaded.
#' * `Asfr` numeric matrix of age specific fertility in 5-year age groups ages 15-19 until 45-49 in rows, and (potentially interpolated) time in columns. Potentially downloaded.
#' * `Exposure_female` numeric matrix of approximated age-specific exposure in 5-year age groups ages 15-19 until 45-49 in rows, and (potentially interpolated) time in columns.
#' * `Bt` births at three time points prior to census corresponding to the midpoints of the cohorts entering ages 0, 1-4, and 5-9.
#' * `SRB` sex ratio at birth at three time points prior to census corresponding to the midpoints of the cohorts entering ages 0, 1-4, and 5-9. Potentially downloaded.
#' * `Age` age groups of the input population counts.
#'
# #' `basepop_single` is used, the return value is a numeric vector with
# #' **single year age groups** where the counts between 0 and 10 are adjusted.
#'
#' @param location UN Pop Division `LocName` or `LocID`
#' @param refDate The reference year for which the reported population pertain
#' (these are the population counts in `Females_five` and
#' \code{Males_five}). Can either be a decimal date, a `Date` class.
#' If \code{nLxDatesIn} or \code{AsfrDatesIn} are not supplied and the
#' corresponding \code{nLxFemale/Male}/\code{AsfrMat} is not supplied,
#' \code{refDate} must be at a minimum 1962.5. This is because we can only
#' fetch WPP data from 1955 onwards, and these minimum date is assumed to be
#' 7.5 years before \code{refDate}, meaning 1955.
#'
#' @param Age integer vector of lower bounds of abridged age groups given in `Females_five` and `Males_five`.
#'
#' @param Females_five A named numeric vector with the population counts for
#' five-year abridged age groups for females in `refDate`. The names of the
#' vector should reflect the age groups. See the example section for some
#' examples.
#'
#' @param nLxFemale A numeric matrix. The female nLx function of two abridged life tables
#' with ages in the rows and time in columns. The earlier date should be at least
#' 7.5 years before the reference date of the "reported" population. The later
#' date should be no earlier than one-half year before the reference date of
#' the "reported" population. If not provided, it's automatically downloaded if
#' `location`, `refDate` and the equivalent population counts
#' `*_five` are provided.
#'
#' @param nLxDatesIn A vector of numeric years (for example, 1986). The dates
#' which pertain to the columns in `nLxFemale` and `nLxMale`. If not
#' provided, the function automatically determines two dates which are 8 years
#' before `refDate` and 0.5 years after `refDate`.
#'
#' @param AsfrMat A numeric matrix. An age-period matrix of age specific
#' fertility rates with age in rows, time in columns. If not provided, the
#' function automatically downloads the ASFR matrix based on the dates in
#' `AsfrDatesIn`.
#'
#' @param AsfrDatesIn A vector of numeric years (for example, 1986). These are
#' the dates which pertain to the columns in `AsfrMat`. If not provided,
#' the function automatically determines two dates which are 8 years before
#' `refDate` and 0.5 before `refDate`.
#'
#' @param ... Arguments passed to `\link{interp}`. In particular, users
#' might be interested in changing the interpolation method for the `nLx*`
#' matrices and the `Asfr` matrix. By default, it's linearly interpolated.
#'
#' @param Males_five A named numeric vector with the population counts for
#' five-year abridged age groups for males in `refDate`. The names of
#' the vector should reflect the age groups. See the example section for
#' some examples.
#'
#' @param nLxMale A numeric matrix. The male nLx function of two abridged life tables
#' with ages in the rows and time in columns. The dates which are represented
#' in the columns are assumed to be the same as `nLxDatesIn`. This
#' argument is only used when `female` is set to `FALSE` and
#' `Males_five` is provided. If `Males_five` is provided and
#' `female` set to `FALSE`, the `nLx` for males is
#' automatically downloaded for the dates in `nLxDatesIn`.
#'
#' @param SRB A numeric. Sex ratio at birth (males / females). Default is set
#' to 1.046. Only a maximum of three values permitted.
#'
#' @param SRBDatesIn A vector of numeric years (for example, 1986). Only a maximum
#' number of three dates allowed. These are
#' the dates which pertain to the values in `SRB`. If not provided,
#' the function automatically determines three dates which are 7.5 years,
#' 2.5 and 0.5 years before `refDate`.
#'
#' @param radix starting point to use in the adjustment of the three first age
#' groups. Default is NULL. If not provided, it is inferred based on the scale of age `1L0`.
#'
#' @param verbose when downloading new data, should the function print details
#' about the download at each step? Defaults to `TRUE`. We recommend the
#' user to set this to `TRUE` at all times because the function needs to
#' make decisions (such as picking the dates for the Asfr and nLx) that the user
#' should be aware of.
#'
#' @export
#' @examples
#'
#' \dontrun{
#'
#' ################ BPE (five year age groups) #####################
#'
#' # Grab population counts for females
#' refDate <- 1986
#' location <- "Brazil"
#' pop_female_single <- fertestr::FetchPopWpp2019(location,
#' refDate,
#' ages = 0:100,
#' sex = "female")
#' pop_female_counts <- single2abridged(setNames(pop_female_single$pop,
#' pop_female_single$ages))
#' pop_male_single <- fertestr::FetchPopWpp2019(location,
#' refDate,
#' ages = 0:100,
#' sex = "male")
#' pop_male_counts <- single2abridged(setNames(pop_male_single$pop,
#' pop_male_single$ages))
#' Age <- names2age(pop_male_counts)
#' # Automatically downloads the nLx, ASFR, and SRB data
#' bpe <- basepop_five(
#' location = location,
#' refDate = refDate,
#' Females_five = pop_female_counts,
#' Males_five = pop_male_counts,
#' Age = Age
#' )
#'
#' # The counts for the first three age groups have been adjusted:
#' bpe$Females_adjusted[1:3]
#' pop_female_counts[1:3]
#'
#' bpe$Males_adjusted[1:3]
#' pop_male_counts[1:3]
#'
#'
#' ################ BPE (for single ages) ############################
#' # blocked out for now, until single age function refactored as
#' # TR: actually, it just needs to be rethought for single ages..
#' # pop_female_single <- setNames(pop_female_single$pop, pop_female_single$ages)
#' #
#' # # Automatically downloads the nLx and ASFR data
#' # bpe_female <- basepop_single(
#' # location = location,
#' # refDate = refDate,
#' # Females_single = pop_female_single
#' # )
#' #
#' # # The counts for the first 10 age groups have been adjusted:
#' # bpe_female[1:10]
#' # pop_female_single[1:10]
#' ################ BPA (five year age groups) #####################
#' # for BPA, smooth counts in advance
#' smoothed_females <- smooth_age_5(Value = pop_female_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' # Note, smooth_age_5() will group infants into the 0-4 age group. So,
#' # we manually stick them back in place.
#' smoothed_females <- c(pop_female_counts[1:2], smoothed_females[-1])
#' smoothed_males <- smooth_age_5(Value = pop_male_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' smoothed_males <- c(smoothed_males[1:2], smoothed_males[-1])
#'
#' # Automatically downloads the nLx, ASFR, and SRB data
#' bpa <- basepop_five(
#' location = location,
#' refDate = refDate,
#' Females_five = smoothed_females,
#' Males_five = smoothed_males,
#' Age = Age
#' )
#'
#' # The counts for the first three age groups have been adjusted:
#' bpa$Females_adjusted[1:3]
#' smoothed_females[1:3]
#' pop_female_counts[1:3]
#'
#' bpa$Males_adjusted[1:3]
#' smoothed_males[1:3]
#' pop_male_counts[1:3]
#'
#' ################ PAS example ###############################
#'
#' # (1) refDate
#' refDate <- 1986.21
#'
#' # (2) Reported population by 5-year age groups and sex in the base year
#' # (Include unknowns).
#'
#' pop_male_counts <- c(11684, 46738, 55639, 37514, 29398, 27187, 27770, 20920, 16973,
#' 14999, 11330, 10415, 6164, 7330, 3882, 3882, 1840, 4200)
#'
#' pop_female_counts <- c(11673, 46693, 55812, 35268, 33672, 31352, 33038, 24029, 16120,
#' 14679, 8831, 9289, 4172, 6174, 2715, 3344, 1455, 4143)
#' Age <- c(0,1, seq(5, 80, by = 5))
#'
#' # (4) Sex ratio at birth (m/f)
#' sex_ratio <- 1.0300
#'
#' # (6) The male and female nLx functions for ages under 1 year, 1 to 4 years, and 5 to 9
#' # years, pertaining to an earlier and later date
#' nLxDatesIn <- c(1977.31, 1986.50)
#'
#' nLxMale <- matrix(c(87732, 304435, 361064, 88451, 310605, 370362),
#' nrow = 3, ncol = 2)
#'
#' nLxFemale <- matrix(c(89842, 314521, 372681, 353053, 340650, 326588,
#' 311481, 295396, 278646, 261260, 241395,217419,
#' 90478, 320755, 382531, 364776, 353538, 340687,
#' 326701, 311573, 295501, 278494, 258748,234587),
#' nrow = 12,
#' ncol = 2)
#'
#' # (7) A set of age-specific fertility rates pertaining to an earlier and later
#' # date
#'
#' asfrmat <- structure(
#' c(0.2, 0.3, 0.3, 0.25, 0.2, 0.15, 0.05, 0.15, 0.2,
#' 0.275, 0.225, 0.175, 0.125, 0.05), .Dim = c(7L, 2L),
#' .Dimnames = list(
#' c("15-19", "20-24", "25-29", "30-34", "35-39", "40-44", "45-49"),
#' c("1977.81", "1985.71")))
#'
#' # for BPA, smooth counts in advance
#' smoothed_females <- smooth_age_5(Value = pop_female_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' smoothed_females <- c(pop_female_counts[1:2], smoothed_females[-1])
#' smoothed_males <- smooth_age_5(Value = pop_male_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' smoothed_males <- c(pop_male_counts[1:2], smoothed_males[-1])
#' ## This is the only number that messes up the whole calculation.
#' ## smooth_age_5 returns the same result as the PASS excel sheet
#' ## except for the age groups 10-15 and 15-19. Here we only use
#' ## age group 15-19. If we plug in manually the correct value,
#' ## we get all results match exactly, otherwise there are
#' ## some differences.
#' smoothed_females[4] <- 34721
#'
#' # For adjusting using BPA for males, we need to specify
#' # female = FALSE with Males and nLxMale.
#' bpa <-
#' basepop_five(
#' refDate = refDate,
#' Males_five = smoothed_males,
#' Females_five = smoothed_females,
#' Age = Age,
#' SRB = sex_ratio,
#' nLxFemale = nLxFemale,
#' nLxMale = nLxMale,
#' nLxDatesIn = nLxDatesIn,
#' AsfrMat = asfrmat,
#' AsfrDatesIn = AsfrDatesIn,
#' radix = 1e5
#' )
#'
#' # See adjustments?
#' pop_male_counts[1:3]
#' bpa$Male_adjusted[1:3]
#'
#' pop_female_counts[1:3]
#' bpa$Female_adjusted[1:3]
#'
#' # For adjustment using BPE, we use exactly the same definitions as above
#' # but use the original inputs
#'
#' bpe <-
#' basepop_five(
#' refDate = refDate,
#' Females_five = pop_female_counts,
#' Males_five = pop_male_counts,
#' SRB = sex_ratio,
#' nLxFemale = nLxFemale,
#' nLxDatesIn = nLxDatesIn,
#' AsfrMat = asfrmat,
#' AsfrDatesIn = AsfrDatesIn
#' )
#'
#' pop_female_counts[1:3]
#' bpe$Females_adjusted[1:3]
#'
#' # basepop_single for single ages
#' # Single ages for males and females
#'
#' # pop_male_counts <-
#' # c(11684, 11473, 11647, 11939, 11680, 10600, 11100, 11157, 11238,
#' # 11544, 7216, 7407, 7461, 7656, 7774, 5709, 5629, 5745, 6056,
#' # 6259, 5303, 5423, 5497, 5547, 5417, 5441, 5466, 5500, 5668, 5694,
#' # 4365, 4252, 4122, 4142, 4039, 3210, 3222, 3258, 3413, 3871, 2684,
#' # 2844, 3052, 3182, 3237, 2263, 2298, 2318, 2257, 2194, 2231, 2172,
#' # 2072, 2008, 1932, 1301, 1262, 1213, 1197, 1191, 1601, 1593, 1490,
#' # 1348, 1299, 568, 745, 843, 801, 925, 806, 883, 796, 725, 672,
#' # 470, 441, 340, 300, 289, 4200)
#' #
#' # pop_female_counts <-
#' # c(11673, 11474, 11670, 11934, 11614, 10603, 11144, 11179, 11269,
#' # 11617, 6772, 6948, 7030, 7211, 7306, 6531, 6443, 6535, 6951,
#' # 7213, 6096, 6234, 6327, 6410, 6285, 6464, 6492, 6549, 6739, 6795,
#' # 5013, 4888, 4735, 4747, 4646, 3040, 3068, 3107, 3246, 3658, 2650,
#' # 2788, 2977, 3108, 3156, 1756, 1784, 1802, 1764, 1724, 1982, 1935,
#' # 1846, 1795, 1731, 863, 850, 825, 819, 816, 1348, 1342, 1246,
#' # 1138, 1101, 391, 520, 585, 560, 659, 670, 750, 686, 634, 604,
#' # 353, 340, 270, 246, 247, 4143)
#' # Age <- 0:80
#' #
#' # smoothed_females <- smooth_age_5(Value = pop_female_counts,
#' # Age = Age,
#' # method = "Arriaga",
#' # OAG = TRUE,
#' # young.tail = "Original")
#' # smoothed_males <- smooth_age_5(Value = pop_male_counts,
#' # Age = Age,
#' # method = "Arriaga",
#' # OAG = TRUE,
#' # young.tail = "Original")
#'
#' # For adjusting using BPA for males, we need to specify
#' # female = FALSE with Males and nLxMale.
#'
#' # This needs work still
#' # bpa_male <-
#' # basepop_single(
#' # refDate = refDate,
#' # Males_single = pop_male_counts,
#' # Females_single = pop_female_counts,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxMale = nLxMale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn
#' # )
#'
#' # See adjustments?
#' # pop_male_counts[1:10]
#' # bpa_male[1:10]
#'
#' # Adjusting the BPA for females requires less arguments
#' # bpa_female <-
#' # basepop_single(
#' # refDate = refDate,
#' # Females_single = pop_female_counts,
#' # SmoothedFemales = smoothed_females,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn
#' # )
#'
#' # pop_female_counts[1:10]
#' # bpa_female[1:10]
#' #
#' # # For adjustment using BPE, we use exactly the same definitions as above
#' # # but remove SmoothedFemales.
#' # bpe_male <-
#' # basepop_single(
#' # refDate = refDate,
#' # Males_single = pop_male_counts,
#' # Females_single = pop_female_counts,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxMale = nLxMale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn,
#' # female = FALSE
#' # )
#'
#' # See adjustments?
#' # pop_male_counts[1:10]
#' # bpa_male[1:10]
#' # bpe_male[1:10]
#'
#' # Adjusting the BPA for females requires less arguments
#' # bpe_female <-
#' # basepop_single(
#' # refDate = refDate,
#' # Females_single = pop_female_counts,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn
#' # )
#' #
#' # pop_female_counts[1:10]
#' # bpa_female[1:10]
#' # bpe_female[1:10]
#'
#' }
#'
#' @references
#' \insertRef{arriaga1994population}{DemoTools}
#' \insertRef{PAS}{DemoTools}
#'
basepop_five <- function(location = NULL,
refDate,
Age = NULL,
Females_five,
Males_five = NULL,
nLxFemale = NULL,
nLxMale = NULL,
nLxDatesIn = NULL,
AsfrMat = NULL,
AsfrDatesIn = NULL,
...,
SRB = NULL,
SRBDatesIn = NULL,
radix = NULL,
verbose = TRUE) {
options(basepop_verbose = verbose)
on.exit(options(basepop_verbose = NULL))
# Ensure census date is numeric.
# "YYYY-MM-DD" input is acceptable
refDate <- dec.date(refDate)
if (!is.null(Age)){
stopifnot(is_abridged(Age))
stopifnot(length(Age) == length(Females_five))
} else {
if (!is.null(names(Females_five))){
Age <- names2age(Females_five)
} else {
if (verbose) {
cat("Assuming age groups are in standard abridged intervals")
}
# last resort = assume in abrided ages!
Age <- inferAgeIntAbr(Females_five)
}
}
if (is.null(nLxDatesIn)) {
# re PJ issue #183 suggested default
nLxDatesIn <- refDate - c(0.5, 7.5)
#nLxDatesIn <- c(abs(8 - refDate), refDate + 0.5)
if (verbose) {
cat(paste0("Assuming the two prior dates for the nLx matrix to be: ", paste0(nLxDatesIn, collapse = ", ")), sep = "\n")
}
}
if (is.null(AsfrDatesIn)) {
# re PJ issue #183 suggested default
AsfrDatesIn <- refDate - c(0.5, 7.5)
#AsfrDatesIn <- abs(c(8, 0.5) - refDate)
if (verbose) {
cat(paste0("Assuming the two prior dates for the Asfr matrix to be: ", paste0(AsfrDatesIn, collapse = ", ")), sep = "\n")
}
}
# ensure vectors named, for purposes of selection
names(Females_five) <- Age
names(Males_five) <- Age
## obtain nLx for males and females
## If these arguments have been specified, they return
## the same thing and don't download the data
nLxFemale <-
downloadnLx(
nLx = nLxFemale,
location = location,
gender = "female",
nLxDatesIn = nLxDatesIn
)
nLxMale <-
downloadnLx(
nLx = nLxMale,
location = location,
gender = "male",
nLxDatesIn = nLxDatesIn
)
if (is.null(radix)) {
# TR: not perfect, but it's a better guess. It would seem the radix
# being pulled before was always 1, whereas the nLx columns was based on 100000
radix <- lt_infer_radix_from_1L0(nLxMale[1,1])
if (verbose) {
cat(paste0("Setting radix to value of lx: ", radix, ". Can be overwritten with the `radix` argument"), sep = "\n")
}
}
AsfrMat <-
downloadAsfr(
Asfrmat = AsfrMat,
location = location,
AsfrDatesIn = AsfrDatesIn
)
DatesOut <- refDate - c(0.5, 2.5, 7.5)
SRBDatesIn <- if (!is.null(SRBDatesIn)) SRBDatesIn else DatesOut
SRB <- downloadSRB(SRB,
location,
DatesOut = SRBDatesIn,
verbose = verbose)
## Check all arguments
AllArgs <- as.list(environment())
ArgsCheck(AllArgs)
lower_bound <- abs(min(nLxDatesIn) - min(DatesOut))
upper_bound <- abs(max(nLxDatesIn) - max(DatesOut))
if (lower_bound > 5 || upper_bound > 5) {
stop("nLxDatesIn implies an extrapolation of > 5 years to achieve the needed reference dates")
}
# Interpolate the gender specific nLx to the requested
# dates out
nLxf <- interp(
nLxFemale,
datesIn = nLxDatesIn,
datesOut = DatesOut,
...
)
nLxm <- interp(
nLxMale,
datesIn = nLxDatesIn,
datesOut = DatesOut,
...
)
lower_bound <- abs(min(AsfrDatesIn) - min(DatesOut))
upper_bound <- abs(max(AsfrDatesIn) - max(DatesOut))
if (lower_bound > 5 || upper_bound > 5) {
stop("AsfrDatesIn implies an extrapolation of > 5 years to achieve the needed reference dates")
}
# Interpolate the asfr to the requested dates.
# This is gender agnostic.
Asfr <- interp(
AsfrMat,
datesIn = AsfrDatesIn,
datesOut = DatesOut,
...
)
# TR: Follows spreadsheet logic, can still be more elegant.
# sometimes character indexing, sometimes position, but still
ages_15_55 <- as.character(seq(15,55,by=5))
ages_20_55 <- ages_15_55[-1]
ages_15_50 <- ages_15_55[-9]
ages_20_50 <- ages_15_55[-c(1,9)]
ages_15_45 <- ages_15_55[-c(8,9)]
ages_20_45 <- ages_15_55[-c(1,8,9)]
ages_15_40 <- ages_15_55[-c(7,8,9)]
FMiddleages <- Females_five[ages_15_55]
Ft_minus_5 <- FMiddleages[ages_20_55] *
nLxf[ages_15_50, 2] / nLxf[ages_20_55, 2]
names(Ft_minus_5) <- ages_15_50
Ft_minus_10 <- Ft_minus_5[ages_20_50] *
nLxf[ages_15_45, 3] / nLxf[ages_20_50, 3]
names(Ft_minus_10) <- ages_15_45
# Now we take some averages to get to midpoints
Ft_minus_.5 <- FMiddleages[ages_15_45] * .9 + Ft_minus_5[ages_15_45] * .1
Ft_minus_2.5 <- FMiddleages[ages_15_45] * .5 + Ft_minus_5[ages_15_45] * .5
Ft_minus_7.5 <- Ft_minus_5[ages_15_45] * .5 + Ft_minus_10[ages_15_45] * .5
# 3 column matrix of sort-of-exposures for ages 15-45, matched to ASFR
fExpos <- cbind(Ft_minus_.5, Ft_minus_2.5, Ft_minus_7.5)
# Calculate births
Bt <- colSums(fExpos * Asfr)
#GenderCounts <- if (male) Males_five else Females_five
Males_five_out <- Males_five
Females_five_out <- Females_five
## Currently, this assumes that there can only be 3 dates.
## We only have 3 age groups to adjust and 3 dates
PF <- 1 / (SRB + 1)
# Age 0
Females_five_out[1] <- Bt[1] * PF[1] * nLxf[1, 1] / radix
Males_five_out[1] <- Bt[1] * (1 - PF[1]) * nLxm[1, 1] / radix
# Age 1-4
Females_five_out[2] <- Bt[2] * PF[2] * 5 *
sum(nLxf[1:2, 2]) / (radix * 5) -
Females_five_out[1]
Males_five_out[2] <- Bt[2] * (1 - PF[2]) * 5 *
sum(nLxm[1:2, 2]) / (radix * 5) -
Males_five_out[1]
# Age 5-9
Females_five_out[3] <- Bt[3] * PF[3] * 5 *
sum(nLxf[1:2,3]) / (radix * 5) *
nLxf[3,2] / sum(nLxf[1:2,2])
Males_five_out[3] <- Bt[3] * (1 - PF[3]) * 5 *
sum(nLxm[1:2,3]) / (radix * 5) *
nLxm[3,2] / sum(nLxm[1:2,2])
# return the important things
list(
Females_adjusted = Females_five_out,
Males_adjusted = Males_five_out,
Females_five = Females_five,
Males_five = Males_five,
nLxf = nLxf,
nLxm = nLxm,
Asfr = Asfr,
Exposure_female = fExpos,
Bt = Bt,
SRB = SRB,
Age = Age
)
}
# #' @rdname basepop_five
# #' @aliases basepop_five
# #' @param Females_single A named numeric vector. Reported population by 1-year age # groups for \code{refDate} for females. The names of the vector should reflect the age # groups. See examples. The method assumes that the last age group is open (for example# , the population ends at '80+' and '100+')
# #' @param Males_single A named numeric vector. Reported population by 1-year age # groups format \code{refDate} for males. The names of the vector should reflect the # age groups. See examples. The method assumes that the last age group is open (for # example, the population ends at '80+' and '100+')
# #'
# #' @export
# #'
# basepop_single <- function(location = NULL,
# refDate,
# Females_single,
# nLxFemale = NULL,
# nLxDatesIn = NULL,
# AsfrMat = NULL,
# AsfrDatesIn = NULL,
# ...,
# female = TRUE,
# SmoothedFemales = NULL,
# Males_single = NULL,
# nLxMale = NULL,
# SRB = 1.05,
# radix = NULL,
# verbose = TRUE) {
#
# stopifnot(
# !is.null(names(Females_single)),
# is_single(as.numeric(names(Females_single)))
# )
#
# Females_abridged <- single2abridged(Females_single)
# males_present <- !is.null(Males_single)
#
# if (males_present) {
# stopifnot(
# !is.null(names(Males_single)),
# is_single(as.numeric(names(Males_single)))
# )
#
# Males_abridged <- single2abridged(Males_single)
# gender_single <- Males_single
# } else {
# Males_abridged <- Males_single
# gender_single <- Females_single
# }
#
# res <-
# basepop_five(
# location = location,
# refDate = refDate,
# Females_five = Females_abridged,
# nLxFemale = nLxFemale,
# nLxDatesIn = nLxDatesIn,
# AsfrMat = AsfrMat,
# AsfrDatesIn = AsfrDatesIn,
# ... = ...,
# female = female,
# SmoothedFemales = SmoothedFemales,
# Males_five = Males_abridged,
# nLxMale = nLxMale,
# SRB = SRB,
# radix = radix
# )
#
# # Since diff always returns a vector of length `length(x) - 1`,
# # the 1 in the end is to reflct the the open ages for 80+ or 100+
# AgeBins1 <- c(diff(as.integer(names(gender_single))), 1)
# AgeBins2 <- c(diff(as.integer(names(res))), 1)
#
# rescaled_res <-
# rescaleAgeGroups(
# Value1 = gender_single,
# AgeInt1 = AgeBins1,
# Value2 = res,
# AgeInt2 = AgeBins2,
# splitfun = graduate_uniform
# )
#
# round(rescaled_res, 3)
# }
#
# TR: modified to assume males and females always given
ArgsCheck <- function(ArgList) {
with(ArgList, {
stopifnot(
is.numeric(Females_five),
is.numeric(Males_five),
is.numeric(SRB),
is.matrix(nLxFemale),
is.matrix(nLxMale),
is.matrix(AsfrMat),
#is.logical(female),
ncol(nLxFemale) == length(nLxDatesIn),
ncol(nLxMale) == length(nLxDatesIn)
# TR no check on ASFRmat dates?
)})
}
lt_infer_radix_from_1L0 <- function(L0){
if (L0 > 1){
radix_check <- L0 %>% as.integer() %>% log10()
is_it_a_radix <- (radix_check - round(radix_check)) == 0
if (!is_it_a_radix){
pow <- L0 %>% round() %>% as.integer() %>% nchar()
the_radix <- 10^pow
} else {
the_radix <- L0
}
} else {
the_radix <- 1
}
the_radix
}
timriffe/DemoTools documentation built on Jan. 28, 2024, 5:13 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions lt_infer_radix_from_1L0 ArgsCheck basepop_five
Documented in basepop_five
#' BPA and BPE methods for adjusting age groups under 10
#' @description Adjust population counts for the age groups 0 to 10
#'
#' @details
#'
#' \code{basepop_five} and \code{basepop_single} can estimate both the BPA and
#' BPE methods. If the user specifies \code{SmoothedFemales}, both
#' \code{basepop_*} functions will return the BPA method.
#' If \code{SmoothedFemales} is left empty, both \code{basepop_*} functions will
#' adjust using the BPE method.
#'
#' For \code{basepop_five}, adjusting the female population counts is the
#' default. For this, only the \code{location}, \code{refDate} and
#' \code{Females_five} are needed. All other arguments are downloaded
#' or set to sensible defaults. For adjusting the male population
#' counts, the user needs to specify the \code{Males_five} population
#' counts and set \code{female = FALSE}.
#'
# # For \code{basepop_single}, the same procedure applies. The only difference
# # is that the vector \code{Males_five} is named \code{Males_single} and accepts
# # a vector of single ages rather than five year abridged age groups. Similarly,
# # the vector for females is \code{Females_single} rather than
# # \code{Females_five} and accepts single age groups.
#'
#' Currently, \code{basepop_five} works only with five year abridged age groups
# #' while \code{basepop_single} works only with single year age groups.
#'
#' The BPE method is used by default. To adjust the counts using
#' the BPA method, the user needs to provide the \code{SmoothedFemales}
#' argument. This is the female population counts passed through
#' a smoothing function such as \code{smooth_age_5}. See the examples
#' section for some examples.
#'
#' @section BPA:
#'
#' Description:
#'
#' The method estimates a smoothed population ages 10 and over and adjusts
#' the population under age 10 using the smoothed population and estimates
#' of fertility and mortality.
#'
#' Based on the smoothed female population counts, it rejuvenates the female
#' "reported" population 20 to 59 years of age for the two 5 year periods prior
#' to the census date to represent the female population in reproductive ages
#' 5 and 10 years earlier. Based on the rejuvenated population and fertility
#' and mortality levels, the method then estimates the male and female births
#' during the two 5 year periods prior to the census date. Next, it projects
#' the two 5-year birth cohorts to the census date. The projected figures
#' represent the adjusted population ages 0 to 4 years and 5 to 9 years
#' at the census date.
#'
#' Advantages:
#'
#' (1) The method adjusts under-10 population to be consistent with fertility
#' and mortality levels and adjusted adult female population.
#'
#' Limitations:
#'
#' (1) BPA assumes a linear change in fertility and mortality during the decade
#' prior to the reference year.
#'
#' (2) The procedure ignores migration, which can lead to misleading results.
#' There are two issues. First, age groups 0-4 and 5-9 are subject to migration,
#' which would affect the comparability of estimated and reported populations
#' in the base year. Second, the estimated size of age groups 0-4 and 5-9 are
#' calculated from numbers of women of reproductive age in the base year
#' rejuvenated to points in the past. With migration, rejuvenated number of
#' women may exceed or be smaller than the number present, and giving
#' birth to children, in the decade prior to the base year.
#'
#' (3) BPA’s smoothing calculations may mask unusual, but real, variations
#' in population age group size. Smoothing irregularities in age structure
#' not attributable to age misreporting will distort estimated births and
#' survived children in the base year.
#'
#' Assumptions:
#'
#' (1) No significant international migration took place within the
#' reference periods for the population, mortality, and fertility input.
#'
#' (2) The data input as the "reported" population is not affected by
#' underenumeration of persons in certain ages, nor by age misreporting.
#'
#' @section BPE:
#'
#' Description:
#'
#' The method adjusts the population under age 10 using the reported population
#' ages 10 and above and estimates of fertility and mortality.
#'
#' The method rejuvenates the reported female population 20 to 59 years of age
#' for the two 5 year periods prior to the census date to represent the female
#' population in reproductive ages 5 and 10 years earlier. Based on the
#' rejuvenated population and fertility and mortality levels, the method then
#' estimates the male and female births during the two 5 year periods prior to
#' the census date. Next, it projects the two 5-year birth cohorts to the
#' census date. The projected figures represent the adjusted population ages
#' 0 to 4 years and 5 to 9 years at the census date.
#'
#' Advantages:
#'
#' (1) The method adjusts the under-10 population to be consistent with
#' fertility and mortality levels and adult female population.
#'
#' Limitations:
#'
#' (1) BPE assumes a linear change in fertility and mortality during the decade
#' prior to the reference year.
#'
#' (2) The procedure ignores migration, which can lead to misleading results.
#' There are two issues. First, age groups 0-4 and 5-9 are subject to
#' migration, which would affect the comparability of estimated and reported
#' populations in the base year. Second, the estimated size of age groups
#' 0-4 and 5-9 are calculated from numbers of women of reproductive age in
#' the base year rejuvenated to points in the past. With migration, rejuvenated
#' number of women may exceed or be smaller than the number present, and
#' giving birth to children, in the decade prior to the base year.
#'
#' (3) The method does not adjust for possible underenumeration and age
#' misreporting errors in the over-10 “reported” population. If the
#' reported population is subject to age-misreporting or age-sex-specific
#' underenumeration, the over-10 population should be smoothed or otherwise
#' corrected prior to use.
#'
#' Assumptions:
#'
#' (1) No significant international migration took place within the reference
#' periods for the population, mortality, and fertility input.
#'
#' (2) The data input as the “reported” population is not affected by
#' underenumeration of persons in certain ages, nor by age misreporting.
#'
#' @return `basepop_five` returns a list with the following elements:
#' *
#' * `Females_adjusted` numeric vector of adjusted population counts for females. Age groups 0, 1-4, and 5-9 are adjusted, while ages 10 and higher are unchanged.
#' * `Males_adjusted` numeric vector of adjusted population counts for males. Age groups 0, 1-4, and 5-9 are adjusted, while ages 10 and higher are unchanged.
#' * `Females_five` numeric vector of female population counts given as input.
#' * `Males_five` numeric vector of male population counts given as input.
#' * `nLxf` numeric matrix of female `nLx`, abridged ages in rows and (potentially interpolated) time in columns. Potentially downloaded.
#' * `nLxm` numeric matrix of male `nLx`, abridged ages in rows and (potentially interpolated) time in columns. Potentially downloaded.
#' * `Asfr` numeric matrix of age specific fertility in 5-year age groups ages 15-19 until 45-49 in rows, and (potentially interpolated) time in columns. Potentially downloaded.
#' * `Exposure_female` numeric matrix of approximated age-specific exposure in 5-year age groups ages 15-19 until 45-49 in rows, and (potentially interpolated) time in columns.
#' * `Bt` births at three time points prior to census corresponding to the midpoints of the cohorts entering ages 0, 1-4, and 5-9.
#' * `SRB` sex ratio at birth at three time points prior to census corresponding to the midpoints of the cohorts entering ages 0, 1-4, and 5-9. Potentially downloaded.
#' * `Age` age groups of the input population counts.
#'
# #' `basepop_single` is used, the return value is a numeric vector with
# #' **single year age groups** where the counts between 0 and 10 are adjusted.
#'
#' @param location UN Pop Division `LocName` or `LocID`
#' @param refDate The reference year for which the reported population pertain
#' (these are the population counts in `Females_five` and
#' \code{Males_five}). Can either be a decimal date, a `Date` class.
#' If \code{nLxDatesIn} or \code{AsfrDatesIn} are not supplied and the
#' corresponding \code{nLxFemale/Male}/\code{AsfrMat} is not supplied,
#' \code{refDate} must be at a minimum 1962.5. This is because we can only
#' fetch WPP data from 1955 onwards, and these minimum date is assumed to be
#' 7.5 years before \code{refDate}, meaning 1955.
#'
#' @param Age integer vector of lower bounds of abridged age groups given in `Females_five` and `Males_five`.
#'
#' @param Females_five A named numeric vector with the population counts for
#' five-year abridged age groups for females in `refDate`. The names of the
#' vector should reflect the age groups. See the example section for some
#' examples.
#'
#' @param nLxFemale A numeric matrix. The female nLx function of two abridged life tables
#' with ages in the rows and time in columns. The earlier date should be at least
#' 7.5 years before the reference date of the "reported" population. The later
#' date should be no earlier than one-half year before the reference date of
#' the "reported" population. If not provided, it's automatically downloaded if
#' `location`, `refDate` and the equivalent population counts
#' `*_five` are provided.
#'
#' @param nLxDatesIn A vector of numeric years (for example, 1986). The dates
#' which pertain to the columns in `nLxFemale` and `nLxMale`. If not
#' provided, the function automatically determines two dates which are 8 years
#' before `refDate` and 0.5 years after `refDate`.
#'
#' @param AsfrMat A numeric matrix. An age-period matrix of age specific
#' fertility rates with age in rows, time in columns. If not provided, the
#' function automatically downloads the ASFR matrix based on the dates in
#' `AsfrDatesIn`.
#'
#' @param AsfrDatesIn A vector of numeric years (for example, 1986). These are
#' the dates which pertain to the columns in `AsfrMat`. If not provided,
#' the function automatically determines two dates which are 8 years before
#' `refDate` and 0.5 before `refDate`.
#'
#' @param ... Arguments passed to `\link{interp}`. In particular, users
#' might be interested in changing the interpolation method for the `nLx*`
#' matrices and the `Asfr` matrix. By default, it's linearly interpolated.
#'
#' @param Males_five A named numeric vector with the population counts for
#' five-year abridged age groups for males in `refDate`. The names of
#' the vector should reflect the age groups. See the example section for
#' some examples.
#'
#' @param nLxMale A numeric matrix. The male nLx function of two abridged life tables
#' with ages in the rows and time in columns. The dates which are represented
#' in the columns are assumed to be the same as `nLxDatesIn`. This
#' argument is only used when `female` is set to `FALSE` and
#' `Males_five` is provided. If `Males_five` is provided and
#' `female` set to `FALSE`, the `nLx` for males is
#' automatically downloaded for the dates in `nLxDatesIn`.
#'
#' @param SRB A numeric. Sex ratio at birth (males / females). Default is set
#' to 1.046. Only a maximum of three values permitted.
#'
#' @param SRBDatesIn A vector of numeric years (for example, 1986). Only a maximum
#' number of three dates allowed. These are
#' the dates which pertain to the values in `SRB`. If not provided,
#' the function automatically determines three dates which are 7.5 years,
#' 2.5 and 0.5 years before `refDate`.
#'
#' @param radix starting point to use in the adjustment of the three first age
#' groups. Default is NULL. If not provided, it is inferred based on the scale of age `1L0`.
#'
#' @param verbose when downloading new data, should the function print details
#' about the download at each step? Defaults to `TRUE`. We recommend the
#' user to set this to `TRUE` at all times because the function needs to
#' make decisions (such as picking the dates for the Asfr and nLx) that the user
#' should be aware of.
#'
#' @export
#' @examples
#'
#' \dontrun{
#'
#' ################ BPE (five year age groups) #####################
#'
#' # Grab population counts for females
#' refDate <- 1986
#' location <- "Brazil"
#' pop_female_single <- fertestr::FetchPopWpp2019(location,
#' refDate,
#' ages = 0:100,
#' sex = "female")
#' pop_female_counts <- single2abridged(setNames(pop_female_single$pop,
#' pop_female_single$ages))
#' pop_male_single <- fertestr::FetchPopWpp2019(location,
#' refDate,
#' ages = 0:100,
#' sex = "male")
#' pop_male_counts <- single2abridged(setNames(pop_male_single$pop,
#' pop_male_single$ages))
#' Age <- names2age(pop_male_counts)
#' # Automatically downloads the nLx, ASFR, and SRB data
#' bpe <- basepop_five(
#' location = location,
#' refDate = refDate,
#' Females_five = pop_female_counts,
#' Males_five = pop_male_counts,
#' Age = Age
#' )
#'
#' # The counts for the first three age groups have been adjusted:
#' bpe$Females_adjusted[1:3]
#' pop_female_counts[1:3]
#'
#' bpe$Males_adjusted[1:3]
#' pop_male_counts[1:3]
#'
#'
#' ################ BPE (for single ages) ############################
#' # blocked out for now, until single age function refactored as
#' # TR: actually, it just needs to be rethought for single ages..
#' # pop_female_single <- setNames(pop_female_single$pop, pop_female_single$ages)
#' #
#' # # Automatically downloads the nLx and ASFR data
#' # bpe_female <- basepop_single(
#' # location = location,
#' # refDate = refDate,
#' # Females_single = pop_female_single
#' # )
#' #
#' # # The counts for the first 10 age groups have been adjusted:
#' # bpe_female[1:10]
#' # pop_female_single[1:10]
#' ################ BPA (five year age groups) #####################
#' # for BPA, smooth counts in advance
#' smoothed_females <- smooth_age_5(Value = pop_female_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' # Note, smooth_age_5() will group infants into the 0-4 age group. So,
#' # we manually stick them back in place.
#' smoothed_females <- c(pop_female_counts[1:2], smoothed_females[-1])
#' smoothed_males <- smooth_age_5(Value = pop_male_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' smoothed_males <- c(smoothed_males[1:2], smoothed_males[-1])
#'
#' # Automatically downloads the nLx, ASFR, and SRB data
#' bpa <- basepop_five(
#' location = location,
#' refDate = refDate,
#' Females_five = smoothed_females,
#' Males_five = smoothed_males,
#' Age = Age
#' )
#'
#' # The counts for the first three age groups have been adjusted:
#' bpa$Females_adjusted[1:3]
#' smoothed_females[1:3]
#' pop_female_counts[1:3]
#'
#' bpa$Males_adjusted[1:3]
#' smoothed_males[1:3]
#' pop_male_counts[1:3]
#'
#' ################ PAS example ###############################
#'
#' # (1) refDate
#' refDate <- 1986.21
#'
#' # (2) Reported population by 5-year age groups and sex in the base year
#' # (Include unknowns).
#'
#' pop_male_counts <- c(11684, 46738, 55639, 37514, 29398, 27187, 27770, 20920, 16973,
#' 14999, 11330, 10415, 6164, 7330, 3882, 3882, 1840, 4200)
#'
#' pop_female_counts <- c(11673, 46693, 55812, 35268, 33672, 31352, 33038, 24029, 16120,
#' 14679, 8831, 9289, 4172, 6174, 2715, 3344, 1455, 4143)
#' Age <- c(0,1, seq(5, 80, by = 5))
#'
#' # (4) Sex ratio at birth (m/f)
#' sex_ratio <- 1.0300
#'
#' # (6) The male and female nLx functions for ages under 1 year, 1 to 4 years, and 5 to 9
#' # years, pertaining to an earlier and later date
#' nLxDatesIn <- c(1977.31, 1986.50)
#'
#' nLxMale <- matrix(c(87732, 304435, 361064, 88451, 310605, 370362),
#' nrow = 3, ncol = 2)
#'
#' nLxFemale <- matrix(c(89842, 314521, 372681, 353053, 340650, 326588,
#' 311481, 295396, 278646, 261260, 241395,217419,
#' 90478, 320755, 382531, 364776, 353538, 340687,
#' 326701, 311573, 295501, 278494, 258748,234587),
#' nrow = 12,
#' ncol = 2)
#'
#' # (7) A set of age-specific fertility rates pertaining to an earlier and later
#' # date
#'
#' asfrmat <- structure(
#' c(0.2, 0.3, 0.3, 0.25, 0.2, 0.15, 0.05, 0.15, 0.2,
#' 0.275, 0.225, 0.175, 0.125, 0.05), .Dim = c(7L, 2L),
#' .Dimnames = list(
#' c("15-19", "20-24", "25-29", "30-34", "35-39", "40-44", "45-49"),
#' c("1977.81", "1985.71")))
#'
#' # for BPA, smooth counts in advance
#' smoothed_females <- smooth_age_5(Value = pop_female_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' smoothed_females <- c(pop_female_counts[1:2], smoothed_females[-1])
#' smoothed_males <- smooth_age_5(Value = pop_male_counts,
#' Age = Age,
#' method = "Arriaga",
#' OAG = TRUE,
#' young.tail = "Original")
#' smoothed_males <- c(pop_male_counts[1:2], smoothed_males[-1])
#' ## This is the only number that messes up the whole calculation.
#' ## smooth_age_5 returns the same result as the PASS excel sheet
#' ## except for the age groups 10-15 and 15-19. Here we only use
#' ## age group 15-19. If we plug in manually the correct value,
#' ## we get all results match exactly, otherwise there are
#' ## some differences.
#' smoothed_females[4] <- 34721
#'
#' # For adjusting using BPA for males, we need to specify
#' # female = FALSE with Males and nLxMale.
#' bpa <-
#' basepop_five(
#' refDate = refDate,
#' Males_five = smoothed_males,
#' Females_five = smoothed_females,
#' Age = Age,
#' SRB = sex_ratio,
#' nLxFemale = nLxFemale,
#' nLxMale = nLxMale,
#' nLxDatesIn = nLxDatesIn,
#' AsfrMat = asfrmat,
#' AsfrDatesIn = AsfrDatesIn,
#' radix = 1e5
#' )
#'
#' # See adjustments?
#' pop_male_counts[1:3]
#' bpa$Male_adjusted[1:3]
#'
#' pop_female_counts[1:3]
#' bpa$Female_adjusted[1:3]
#'
#' # For adjustment using BPE, we use exactly the same definitions as above
#' # but use the original inputs
#'
#' bpe <-
#' basepop_five(
#' refDate = refDate,
#' Females_five = pop_female_counts,
#' Males_five = pop_male_counts,
#' SRB = sex_ratio,
#' nLxFemale = nLxFemale,
#' nLxDatesIn = nLxDatesIn,
#' AsfrMat = asfrmat,
#' AsfrDatesIn = AsfrDatesIn
#' )
#'
#' pop_female_counts[1:3]
#' bpe$Females_adjusted[1:3]
#'
#' # basepop_single for single ages
#' # Single ages for males and females
#'
#' # pop_male_counts <-
#' # c(11684, 11473, 11647, 11939, 11680, 10600, 11100, 11157, 11238,
#' # 11544, 7216, 7407, 7461, 7656, 7774, 5709, 5629, 5745, 6056,
#' # 6259, 5303, 5423, 5497, 5547, 5417, 5441, 5466, 5500, 5668, 5694,
#' # 4365, 4252, 4122, 4142, 4039, 3210, 3222, 3258, 3413, 3871, 2684,
#' # 2844, 3052, 3182, 3237, 2263, 2298, 2318, 2257, 2194, 2231, 2172,
#' # 2072, 2008, 1932, 1301, 1262, 1213, 1197, 1191, 1601, 1593, 1490,
#' # 1348, 1299, 568, 745, 843, 801, 925, 806, 883, 796, 725, 672,
#' # 470, 441, 340, 300, 289, 4200)
#' #
#' # pop_female_counts <-
#' # c(11673, 11474, 11670, 11934, 11614, 10603, 11144, 11179, 11269,
#' # 11617, 6772, 6948, 7030, 7211, 7306, 6531, 6443, 6535, 6951,
#' # 7213, 6096, 6234, 6327, 6410, 6285, 6464, 6492, 6549, 6739, 6795,
#' # 5013, 4888, 4735, 4747, 4646, 3040, 3068, 3107, 3246, 3658, 2650,
#' # 2788, 2977, 3108, 3156, 1756, 1784, 1802, 1764, 1724, 1982, 1935,
#' # 1846, 1795, 1731, 863, 850, 825, 819, 816, 1348, 1342, 1246,
#' # 1138, 1101, 391, 520, 585, 560, 659, 670, 750, 686, 634, 604,
#' # 353, 340, 270, 246, 247, 4143)
#' # Age <- 0:80
#' #
#' # smoothed_females <- smooth_age_5(Value = pop_female_counts,
#' # Age = Age,
#' # method = "Arriaga",
#' # OAG = TRUE,
#' # young.tail = "Original")
#' # smoothed_males <- smooth_age_5(Value = pop_male_counts,
#' # Age = Age,
#' # method = "Arriaga",
#' # OAG = TRUE,
#' # young.tail = "Original")
#'
#' # For adjusting using BPA for males, we need to specify
#' # female = FALSE with Males and nLxMale.
#'
#' # This needs work still
#' # bpa_male <-
#' # basepop_single(
#' # refDate = refDate,
#' # Males_single = pop_male_counts,
#' # Females_single = pop_female_counts,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxMale = nLxMale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn
#' # )
#'
#' # See adjustments?
#' # pop_male_counts[1:10]
#' # bpa_male[1:10]
#'
#' # Adjusting the BPA for females requires less arguments
#' # bpa_female <-
#' # basepop_single(
#' # refDate = refDate,
#' # Females_single = pop_female_counts,
#' # SmoothedFemales = smoothed_females,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn
#' # )
#'
#' # pop_female_counts[1:10]
#' # bpa_female[1:10]
#' #
#' # # For adjustment using BPE, we use exactly the same definitions as above
#' # # but remove SmoothedFemales.
#' # bpe_male <-
#' # basepop_single(
#' # refDate = refDate,
#' # Males_single = pop_male_counts,
#' # Females_single = pop_female_counts,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxMale = nLxMale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn,
#' # female = FALSE
#' # )
#'
#' # See adjustments?
#' # pop_male_counts[1:10]
#' # bpa_male[1:10]
#' # bpe_male[1:10]
#'
#' # Adjusting the BPA for females requires less arguments
#' # bpe_female <-
#' # basepop_single(
#' # refDate = refDate,
#' # Females_single = pop_female_counts,
#' # SRB = sex_ratio,
#' # nLxFemale = nLxFemale,
#' # nLxDatesIn = nLxDatesIn,
#' # AsfrMat = asfrmat,
#' # AsfrDatesIn = AsfrDatesIn
#' # )
#' #
#' # pop_female_counts[1:10]
#' # bpa_female[1:10]
#' # bpe_female[1:10]
#'
#' }
#'
#' @references
#' \insertRef{arriaga1994population}{DemoTools}
#' \insertRef{PAS}{DemoTools}
#'
basepop_five <- function(location = NULL,
refDate,
Age = NULL,
Females_five,
Males_five = NULL,
nLxFemale = NULL,
nLxMale = NULL,
nLxDatesIn = NULL,
AsfrMat = NULL,
AsfrDatesIn = NULL,
...,
SRB = NULL,
SRBDatesIn = NULL,
radix = NULL,
verbose = TRUE) {
options(basepop_verbose = verbose)
on.exit(options(basepop_verbose = NULL))
# Ensure census date is numeric.
# "YYYY-MM-DD" input is acceptable
refDate <- dec.date(refDate)
if (!is.null(Age)){
stopifnot(is_abridged(Age))
stopifnot(length(Age) == length(Females_five))
} else {
if (!is.null(names(Females_five))){
Age <- names2age(Females_five)
} else {
if (verbose) {
cat("Assuming age groups are in standard abridged intervals")
}
# last resort = assume in abrided ages!
Age <- inferAgeIntAbr(Females_five)
}
}
if (is.null(nLxDatesIn)) {
# re PJ issue #183 suggested default
nLxDatesIn <- refDate - c(0.5, 7.5)
#nLxDatesIn <- c(abs(8 - refDate), refDate + 0.5)
if (verbose) {
cat(paste0("Assuming the two prior dates for the nLx matrix to be: ", paste0(nLxDatesIn, collapse = ", ")), sep = "\n")
}
}
if (is.null(AsfrDatesIn)) {
# re PJ issue #183 suggested default
AsfrDatesIn <- refDate - c(0.5, 7.5)
#AsfrDatesIn <- abs(c(8, 0.5) - refDate)
if (verbose) {
cat(paste0("Assuming the two prior dates for the Asfr matrix to be: ", paste0(AsfrDatesIn, collapse = ", ")), sep = "\n")
}
}
# ensure vectors named, for purposes of selection
names(Females_five) <- Age
names(Males_five) <- Age
## obtain nLx for males and females
## If these arguments have been specified, they return
## the same thing and don't download the data
nLxFemale <-
downloadnLx(
nLx = nLxFemale,
location = location,
gender = "female",
nLxDatesIn = nLxDatesIn
)
nLxMale <-
downloadnLx(
nLx = nLxMale,
location = location,
gender = "male",
nLxDatesIn = nLxDatesIn
)
if (is.null(radix)) {
# TR: not perfect, but it's a better guess. It would seem the radix
# being pulled before was always 1, whereas the nLx columns was based on 100000
radix <- lt_infer_radix_from_1L0(nLxMale[1,1])
if (verbose) {
cat(paste0("Setting radix to value of lx: ", radix, ". Can be overwritten with the `radix` argument"), sep = "\n")
}
}
AsfrMat <-
downloadAsfr(
Asfrmat = AsfrMat,
location = location,
AsfrDatesIn = AsfrDatesIn
)
DatesOut <- refDate - c(0.5, 2.5, 7.5)
SRBDatesIn <- if (!is.null(SRBDatesIn)) SRBDatesIn else DatesOut
SRB <- downloadSRB(SRB,
location,
DatesOut = SRBDatesIn,
verbose = verbose)
## Check all arguments
AllArgs <- as.list(environment())
ArgsCheck(AllArgs)
lower_bound <- abs(min(nLxDatesIn) - min(DatesOut))
upper_bound <- abs(max(nLxDatesIn) - max(DatesOut))
if (lower_bound > 5 || upper_bound > 5) {
stop("nLxDatesIn implies an extrapolation of > 5 years to achieve the needed reference dates")
}
# Interpolate the gender specific nLx to the requested
# dates out
nLxf <- interp(
nLxFemale,
datesIn = nLxDatesIn,
datesOut = DatesOut,
...
)
nLxm <- interp(
nLxMale,
datesIn = nLxDatesIn,
datesOut = DatesOut,
...
)
lower_bound <- abs(min(AsfrDatesIn) - min(DatesOut))
upper_bound <- abs(max(AsfrDatesIn) - max(DatesOut))
if (lower_bound > 5 || upper_bound > 5) {
stop("AsfrDatesIn implies an extrapolation of > 5 years to achieve the needed reference dates")
}
# Interpolate the asfr to the requested dates.
# This is gender agnostic.
Asfr <- interp(
AsfrMat,
datesIn = AsfrDatesIn,
datesOut = DatesOut,
...
)
# TR: Follows spreadsheet logic, can still be more elegant.
# sometimes character indexing, sometimes position, but still
ages_15_55 <- as.character(seq(15,55,by=5))
ages_20_55 <- ages_15_55[-1]
ages_15_50 <- ages_15_55[-9]
ages_20_50 <- ages_15_55[-c(1,9)]
ages_15_45 <- ages_15_55[-c(8,9)]
ages_20_45 <- ages_15_55[-c(1,8,9)]
ages_15_40 <- ages_15_55[-c(7,8,9)]
FMiddleages <- Females_five[ages_15_55]
Ft_minus_5 <- FMiddleages[ages_20_55] *
nLxf[ages_15_50, 2] / nLxf[ages_20_55, 2]
names(Ft_minus_5) <- ages_15_50
Ft_minus_10 <- Ft_minus_5[ages_20_50] *
nLxf[ages_15_45, 3] / nLxf[ages_20_50, 3]
names(Ft_minus_10) <- ages_15_45
# Now we take some averages to get to midpoints
Ft_minus_.5 <- FMiddleages[ages_15_45] * .9 + Ft_minus_5[ages_15_45] * .1
Ft_minus_2.5 <- FMiddleages[ages_15_45] * .5 + Ft_minus_5[ages_15_45] * .5
Ft_minus_7.5 <- Ft_minus_5[ages_15_45] * .5 + Ft_minus_10[ages_15_45] * .5
# 3 column matrix of sort-of-exposures for ages 15-45, matched to ASFR
fExpos <- cbind(Ft_minus_.5, Ft_minus_2.5, Ft_minus_7.5)
# Calculate births
Bt <- colSums(fExpos * Asfr)
#GenderCounts <- if (male) Males_five else Females_five
Males_five_out <- Males_five
Females_five_out <- Females_five
## Currently, this assumes that there can only be 3 dates.
## We only have 3 age groups to adjust and 3 dates
PF <- 1 / (SRB + 1)
# Age 0
Females_five_out[1] <- Bt[1] * PF[1] * nLxf[1, 1] / radix
Males_five_out[1] <- Bt[1] * (1 - PF[1]) * nLxm[1, 1] / radix
# Age 1-4
Females_five_out[2] <- Bt[2] * PF[2] * 5 *
sum(nLxf[1:2, 2]) / (radix * 5) -
Females_five_out[1]
Males_five_out[2] <- Bt[2] * (1 - PF[2]) * 5 *
sum(nLxm[1:2, 2]) / (radix * 5) -
Males_five_out[1]
# Age 5-9
Females_five_out[3] <- Bt[3] * PF[3] * 5 *
sum(nLxf[1:2,3]) / (radix * 5) *
nLxf[3,2] / sum(nLxf[1:2,2])
Males_five_out[3] <- Bt[3] * (1 - PF[3]) * 5 *
sum(nLxm[1:2,3]) / (radix * 5) *
nLxm[3,2] / sum(nLxm[1:2,2])
# return the important things
list(
Females_adjusted = Females_five_out,
Males_adjusted = Males_five_out,
Females_five = Females_five,
Males_five = Males_five,
nLxf = nLxf,
nLxm = nLxm,
Asfr = Asfr,
Exposure_female = fExpos,
Bt = Bt,
SRB = SRB,
Age = Age
)
}
# #' @rdname basepop_five
# #' @aliases basepop_five
# #' @param Females_single A named numeric vector. Reported population by 1-year age # groups for \code{refDate} for females. The names of the vector should reflect the age # groups. See examples. The method assumes that the last age group is open (for example# , the population ends at '80+' and '100+')
# #' @param Males_single A named numeric vector. Reported population by 1-year age # groups format \code{refDate} for males. The names of the vector should reflect the # age groups. See examples. The method assumes that the last age group is open (for # example, the population ends at '80+' and '100+')
# #'
# #' @export
# #'
# basepop_single <- function(location = NULL,
# refDate,
# Females_single,
# nLxFemale = NULL,
# nLxDatesIn = NULL,
# AsfrMat = NULL,
# AsfrDatesIn = NULL,
# ...,
# female = TRUE,
# SmoothedFemales = NULL,
# Males_single = NULL,
# nLxMale = NULL,
# SRB = 1.05,
# radix = NULL,
# verbose = TRUE) {
#
# stopifnot(
# !is.null(names(Females_single)),
# is_single(as.numeric(names(Females_single)))
# )
#
# Females_abridged <- single2abridged(Females_single)
# males_present <- !is.null(Males_single)
#
# if (males_present) {
# stopifnot(
# !is.null(names(Males_single)),
# is_single(as.numeric(names(Males_single)))
# )
#
# Males_abridged <- single2abridged(Males_single)
# gender_single <- Males_single
# } else {
# Males_abridged <- Males_single
# gender_single <- Females_single
# }
#
# res <-
# basepop_five(
# location = location,
# refDate = refDate,
# Females_five = Females_abridged,
# nLxFemale = nLxFemale,
# nLxDatesIn = nLxDatesIn,
# AsfrMat = AsfrMat,
# AsfrDatesIn = AsfrDatesIn,
# ... = ...,
# female = female,
# SmoothedFemales = SmoothedFemales,
# Males_five = Males_abridged,
# nLxMale = nLxMale,
# SRB = SRB,
# radix = radix
# )
#
# # Since diff always returns a vector of length `length(x) - 1`,
# # the 1 in the end is to reflct the the open ages for 80+ or 100+
# AgeBins1 <- c(diff(as.integer(names(gender_single))), 1)
# AgeBins2 <- c(diff(as.integer(names(res))), 1)
#
# rescaled_res <-
# rescaleAgeGroups(
# Value1 = gender_single,
# AgeInt1 = AgeBins1,
# Value2 = res,
# AgeInt2 = AgeBins2,
# splitfun = graduate_uniform
# )
#
# round(rescaled_res, 3)
# }
#
# TR: modified to assume males and females always given
ArgsCheck <- function(ArgList) {
with(ArgList, {
stopifnot(
is.numeric(Females_five),
is.numeric(Males_five),
is.numeric(SRB),
is.matrix(nLxFemale),
is.matrix(nLxMale),
is.matrix(AsfrMat),
#is.logical(female),
ncol(nLxFemale) == length(nLxDatesIn),
ncol(nLxMale) == length(nLxDatesIn)
# TR no check on ASFRmat dates?
)})
}
lt_infer_radix_from_1L0 <- function(L0){
if (L0 > 1){
radix_check <- L0 %>% as.integer() %>% log10()
is_it_a_radix <- (radix_check - round(radix_check)) == 0
if (!is_it_a_radix){
pow <- L0 %>% round() %>% as.integer() %>% nchar()
the_radix <- 10^pow
} else {
the_radix <- L0
}
} else {
the_radix <- 1
}
the_radix
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.