Skip to contents

Calculate target confidence intervals by zone group and region

calc_target_ci.Rd

Calculates confidence intervals for target (reference) data by zone group and region, joining PUMS data and crosswalks, summarizing by group, and saving outputs. Use for reporting fit and uncertainty of reference targets in weighting workflows.

Usage

calc_target_ci(
  pums_tabbed,
  zone_group_crosswalk,
  puma_client_zone_xwalk,
  ci_level = 0.9,
  run_label,
  settings
)

Arguments

pums_tabbed

data.table. Tabulated PUMS data. Required columns:

  • zone_group <character/integer> — zone group identifier

  • variable — variable name

  • total — total value Other columns as needed for summarization. Rows: one per group-variable. Keys: none required. Modified by reference: no (returns copy).

zone_group_crosswalk

data.table. Zone group crosswalk. Required columns:

  • client_zone_id <character/integer> — client zone identifier

  • zone_group <character/integer> — zone group identifier

  • zone_group_label — zone group label Rows: one per client zone. Keys: none required. Modified by reference: no (read-only).

puma_client_zone_xwalk

data.table. PUMA to client zone crosswalk. Required columns:

  • puma <character/integer> — PUMA code

  • client_zone_id <character/integer> — client zone identifier Rows: one per PUMA-client zone. Keys: none required. Modified by reference: no (read-only).

ci_level

numeric(1). Confidence interval level (e.g., 0.9). Default: 0.9.

run_label

character(1). Run label for output files.

settings

list. Must include:

  • outputs_dir — output directory

  • working_dir — working directory

  • targets — named list of target variables

  • weight_dow_groups — day-of-week groups

  • transit_target_moe — margin of error for transit targets (optional)

Value

list with two data.tables:

  • zone_group: target confidence intervals by zone group. Columns:

    • variable — variable name

    • day_group — day group label

    • lower — lower confidence bound

    • total — total value

    • upper — upper confidence bound

    • Additional columns as needed for reporting

  • region: target confidence intervals by region. Same columns as above.

Details

  • Checks if outputs are up-to-date; re-runs if needed.

  • Creates PUMA-to-zone group crosswalk using client zone and zone group crosswalks.

  • Summarizes PUMS data by zone group and region using helper functions.

  • Drops outside-region zone groups and transit boardings at zone group level.

  • Adds external transit targets to region summary if available.

  • Replicates total targets for each day group, creating dummy DOW targets.

  • Saves outputs as CSV files in the specified output directory.

  • Returns a list of data.tables for zone group and region.

Settings

  • outputs_dir (direct): Output directory. Default from settings.

  • working_dir (direct): Working directory. Default from settings.

  • targets (direct): Named list of target variables. Default from settings.

  • weight_dow_groups (direct): Day-of-week groups. Default from settings.

  • transit_target_moe (direct): Margin of error for transit targets. Default from settings.

See also

  • summarize_pums

  • scripts/weighting/target_fit.R

Examples

## Not run:
pums_tabbed <- data.table(
  zone_group = 1,
  variable = "h_total",
  total = 100
)
zone_group_crosswalk <- data.table(
  client_zone_id = 1,
  zone_group = 1,
  zone_group_label = "A"
)
puma_client_zone_xwalk <- data.table(
  puma = "P1",
  client_zone_id = 1
)
ci_level <- 0.9
run_label <- "test"
settings <- list(
  outputs_dir = "out",
  working_dir = "work",
  targets = list(),
  weight_dow_groups = list(Mon = 1),
  transit_target_moe = 0.05
)
calc_target_ci(
  pums_tabbed,
  zone_group_crosswalk,
  puma_client_zone_xwalk,
  ci_level,
  run_label,
  settings
)
#> Re-running calc_target_ci because inputs are newer than outputs.
#> Error in summarize_pums(pums_wide = pums_tabbed, group_col = "zone_group",     puma_zone_group_xwalk = puma_zone_group_xwalk, ci_level = ci_level,     settings = settings): pums_wide missing columns: puma_id, hh_id, person_id, WGTP, PWGTP, h_total, p_total
## End(Not run)