Disaggregate per-140 counts to per-158 counts (uniform-density) — morie_tps_disaggregate_140_to

Cake-cutting in the FORWARD direction. Given a data.frame of per- historical-hood counts (one row per hood_140, one or more numeric count columns), splits each 140's count across its 158-children in proportion to pct_140_in_158 / 100. For 1:1 hoods the count is passed through unchanged. For splits the count is partitioned (e.g. 140-75 Church-Yonge Corridor's 100 incidents become 59.24 in 158-168 Downtown Yonge East and 40.76 in 158-167 Church-Wellesley).

Usage

morie_tps_disaggregate_140_to_158(
  df,
  hood_140_col = "HOOD_140",
  count_cols = NULL,
  crosswalk = NULL
)

Arguments

df: A data.frame keyed on a 140-hood column.
hood_140_col: Name of the 140-hood column. Default "HOOD_140".
count_cols: Character vector of numeric count columns to disaggregate. Default: every numeric column in df except hood_140_col.
crosswalk: Optional pre-loaded crosswalk; defaults to morie_to_hood_crosswalk().

Value

A data.frame with columns hood_158, name_158, hood_140, all chosen count_cols (now per-158 fractional counts), and pct_140_in_158 (the cake-cut weight applied).

Details

This assumes UNIFORM SPATIAL DENSITY of the underlying events inside each 140-hood – which is the best you can do without per- incident lat/lon. If you have lat/lon, prefer re-binning from points (via sf::st_join against morie_to_neighbourhoods("158", offline = FALSE)) over this uniform-density approximation.

Examples

df <- data.frame(HOOD_140 = c("075", "001"),
                 incidents = c(100, 42))
morie_tps_disaggregate_140_to_158(df)
#>   hood_158               name_158 hood_140 incidents pct_140_in_158
#> 1      168    Downtown Yonge East      075   59.2433        59.2433
#> 2      167       Church-Wellesley      075   40.7567        40.7567
#> 3      001 West Humber-Clairville      001   42.0000       100.0000