Skip to contents

Reshape Gleaning Catch Data from Wide to Long Format

reshape_gleaning_catch.Rd

Reshapes the group_catch section of the Zanzibar/Pemba intertidal gleaning KoboToolbox survey into a tidy long format. Unlike the boat-fishery surveys (see reshape_catch_data_v2()), group_catch is a single (non-repeated) group per submission, so there is no n_catch index. The wide-to-long work instead unpacks three layered structures:

Reshapes the group_catch section of the Zanzibar/Pemba intertidal gleaning KoboToolbox survey into a tidy long format. Unlike the boat-fishery surveys (see reshape_catch_data_v2()), group_catch is a single (non-repeated) group per submission, so there is no n_catch index. The wide-to-long work instead unpacks three layered structures:

Usage

reshape_gleaning_catch(data = NULL)

reshape_gleaning_catch(data = NULL)

Arguments

data

A data frame of the gleaning survey export. Catch columns are expected with the raw Kobo prefixes (group_catch/..., including the nested group_catch/group_length_gastropods/... / group_catch/group_length_bivalves/...).

Value

A long data frame with one row per submission x shell group x size class (plus a single context row for submissions without shell detail).

A long data frame with one row per submission x shell group x size class (plus a single context row for submissions without shell detail).

Details

  1. Counting method. counting_method (backet / plastic_bag) populates two mutually exclusive sets of columns (bucket_* vs plastic_*). These are coalesced into one unified container block: container_type, container_size, container_size_kg, catch_fraction, unit_weight_kg, n_containers.

  2. Shell group + species. shell_group (bivalves / gastropod / both / others) drives which species multi-selects (Group_Bivalves, Group_Gastropod) and which length block apply. Species stay as a space-separated code list per group (no per-species quantities exist in the instrument), with a derived n_species token count.

  3. Length frequency. Individual counts by size class (<5, 5-15, >15 cm) are recorded once per shell GROUP, not per species. Each present group emits one row per size class.

Output grain: one row per submission_id x group x size_class. A submission recorded under bivalves or gastropod yields 3 rows; both yields 6; others or a missing shell_group (no shell detail) is preserved as a single context row with group/size_class/n_individuals = NA so no submission is silently dropped. All three size classes are retained per present group (NA = not recorded, 0 = a recorded zero) so the size-frequency distribution stays explicit and complete.

  1. Counting method. counting_method (backet / plastic_bag) populates two mutually exclusive sets of columns (bucket_* vs plastic_*). These are coalesced into one unified container block: container_type, container_size, container_size_kg, catch_fraction, unit_weight_kg, n_containers.

  2. Shell group + species. shell_group (bivalves / gastropod / both / others) drives which species multi-selects (Group_Bivalves, Group_Gastropod) and which length block apply. Species stay as a space-separated code list per group (no per-species quantities exist in the instrument), with a derived n_species token count.

  3. Length frequency. Individual counts by size class (<5, 5-15, >15 cm) are recorded once per shell GROUP, not per species. Each present group emits one row per size class.

Output grain: one row per submission_id x group x size_class. A submission recorded under bivalves or gastropod yields 3 rows; both yields 6; others or a missing shell_group (no shell detail) is preserved as a single context row with group/size_class/n_individuals = NA so no submission is silently dropped. All three size classes are retained per present group (NA = not recorded, 0 = a recorded zero) so the size-frequency distribution stays explicit and complete.

Examples

if (FALSE) { # \dontrun{
gleaning_long <- reshape_gleaning_catch(gleaning)

# Size-frequency by shell group across all submissions
gleaning_long |>
  dplyr::filter(!is.na(n_individuals)) |>
  dplyr::group_by(group, size_class) |>
  dplyr::summarise(total = sum(n_individuals), .groups = "drop")
} # }

if (FALSE) { # \dontrun{
gleaning_long <- reshape_gleaning_catch(gleaning)

# Size-frequency by shell group across all submissions
gleaning_long |>
  dplyr::filter(!is.na(n_individuals)) |>
  dplyr::group_by(group, size_class) |>
  dplyr::summarise(total = sum(n_individuals), .groups = "drop")
} # }