Core function that uses the cohort component method and matrix algebra to project population development (for more details, see here). The function can be used for different spatial levels (e.g., cantons, municipalities) and for one scenario at a time.
This function provides projections in a raw version in which
key information is missing (e.g., which age groups the rows represent).
To conveniently obtain an enriched, more informative output,
use the wrapper function propop::propop() (which internally uses
propop::project_raw()).
The parameters and starting populations for different spatial levels
can be obtained from the Swiss Federal Statistical Office (FSO).
For instructions on how to download this information from
STAT-TAB,
see
vignette("prepare_data", package = "propop").
The projection parameters need to be passed on as a single data frame to
project_raw with (with the parameters as columns). The column types, names,
and factor levels need to match those specified below.
The 'cohort component method' is implemented with matrices to enhance performance and enable efficient code execution. In a nutshell, the starting population ('n') is multiplied by the survival rate to obtain the people which transition into the projected next year (year + 1). Then, the absolute number of people immigrating from other cantons and other countries is added to the "surviving" population. This results in the starting population for projecting the next year. Newborn children are added separately to the new starting population of each year.
The starting population is clustered in 404 groups: 101 age groups times two nationalities times 2 genders. The survival rate is calculated in the function 'create_transition_matrix()' resulting in the matrix 'L'. We use the rates for mortality, emigration towards countries outside Switzerland, emigration to other cantons and the rate for the acquisition of the Swiss citizenship by the foreign population to calculate survival rates.
For the optional distribution of the population between subregions within the canton, we add the absolute migration balance (German = 'saldo') (immigration + emigration) afterwards.
Steps in this function:
Checks: Checking input data and parameter settings for correct formats.
Data preparation: Preparing vectors, for example, for the projection time frame and creation of empty vectors to be filled with data later on.
Loop over years for calculating the projections
Subsetting parameters: Depending on the selected projection year and on the demographic unit, the parameters for mortality, fertility, acquisition of Swiss citizenship as well as migration parameters are subset by demographic group.
Create matrices: Matrices are built for the survival rate, mortality, fertility and for calculating the number of newborn babies.
Creating vectors: Vectors are built for mortality and migration parameters.
Projection: The transition matrix 'L' is multiplied by the starting population for the next year. Migrating people are added in absolute numbers. People that are 100 years old and older are clustered into one age group (age = 100+). The newborn babies are added to the resulting starting population for the next projection year taking into account their survival rates.
Aggregating the data: All projected years are aggregated into one data frame. The function 'propop()', in which this function is contained, automatically adds relevant meta data to the results.
Usage
project_raw(
parameters,
year_first,
year_last,
age_groups = 101,
fert_first = 16,
fert_last = 50,
share_born_female = 100/205,
n,
subregional
)Arguments
- parameters
data frame containing the FSO rates and numbers to run the projection for a specific spatial level (e.g., canton, municipality).
year: projection year.spatial_unit: ID of spatial entity (e.g., canton, municipality) for which to run the projections.scen: projection scenario, used to subset data frames with multiple scenarios (r = reference, l = low growth, h = high growth scenario).nat: nationality (ch = Swiss; int = foreign / international).sex: sex (f = female, m = male).age: age classes; typically ranging from 0 to 100 (incl. >100).birthrate: numeric, total number of live human births per 1,000 inhabitants.int_mothersproportion of children with Swiss nationality born to non-Swiss mothers.mor: prospective mortality rate (probability of death).acq: rate of acquisition of Swiss citizenship.emi_int: rate of people emigrating abroad.emi_nat: rate of people emigrating to other cantons.imm_int_n: number of people immigrating from abroad.imm_nat_n: number of people immigrating from other cantons.mig_sub: within canton net migration. Useful to account for movements between different subregions (e.g., municipalities). This argument is optional.
- year_first
numeric, first year to be projected.
- year_last
numeric, last year to be projected.
- age_groups
numeric, number of age classes. Creates a vector with 1-year age classes running from
0to (age_groups- 1). Defaults to101(FSO standard number of age groups).- fert_first
numeric, first year of female fertility. Defaults to 16 (FSO standard value).
- fert_last
numeric, last year of female fertility. Defaults to 50 (FSO standard value).
numeric, fraction of female babies. Defaults to 100 / 205 (FSO standard value).
- n
number of people per demographic group and year; should be the year before
year_first. Typically extracted from data frame created withpropop::get_population().- subregional
boolean, TRUE indicates that subregional migration patterns (e.g., movement between municipalities within a canton) are part of the projection.
Value
Returns an unformatted and unlabeled data frame. It includes the number of people for each demographic group per year (starting year and projected years. The number of rows corresponds to the product of years and demographic groups (e.g., nationality (2) X sex (2) X age groups (101) = 404). Variables included in the output:
- scen
projected scenarios.
- spatial_unit
spatial unit for which the projection was run (e.g., canton, district, municipality).
- N
number of people per demographic group.
- IMM_INT
number of immigrants from other countries.
- MIG_NAT
number of people migrating from / to other superordinate spatial units (typically cantons).
- MIG_SUB
number of migrants within the superordinate spatial unit (typically a canton).
- MOR
number of deaths (among people older than 0).
- EMI_INT
number of emigrants to other countries.
- ACQ
number of foreigners who acquire Swiss citizenship (naturalisations).
- BIRTHS
number of births.
Examples
# load package data
data(fso_parameters)
data(fso_population)
# run projection
project_raw(
parameters = fso_parameters |>
dplyr::filter(scen == "reference"),
year_first = 2025,
year_last = 2026,
n = fso_population |> dplyr::pull(n),
subregional = FALSE
) |>
head(10)
#> Running projection for: Aargau (Scenario: reference)
#> ✔ Year: 2025
#> ✔ Year: 2026
#> scen spatial_unit N BIRTHS MOR EMI_INT EMI_NAT IMM_INT
#> 1 reference Aargau 2371 2467.882 8.0133597 4.007841 22.04313 5
#> 2 reference Aargau 2542 0.000 0.9568541 10.513014 48.73828 17
#> 3 reference Aargau 2891 0.000 1.0356512 11.403412 55.98247 16
#> 4 reference Aargau 2766 0.000 1.0947324 12.254949 54.58786 16
#> 5 reference Aargau 2794 0.000 0.0000000 11.310174 44.29472 15
#> 6 reference Aargau 2782 0.000 0.0000000 9.963404 38.86175 14
#> 7 reference Aargau 2787 0.000 0.0000000 9.839934 33.45911 14
#> 8 reference Aargau 2726 0.000 0.0000000 8.912826 28.72282 13
#> 9 reference Aargau 2837 0.000 0.0000000 7.744566 24.20143 12
#> 10 reference Aargau 2866 0.000 0.0000000 8.224463 21.58673 11
#> IMM_NAT ACQ MIG_SUB
#> 1 51 0 0
#> 2 94 0 0
#> 3 84 0 0
#> 4 73 0 0
#> 5 64 0 0
#> 6 50 0 0
#> 7 41 0 0
#> 8 34 0 0
#> 9 30 0 0
#> 10 26 0 0