Possible solution for different reporting units for discards

analysis/x-catch-discard-units.R

@@ -0,0 +1,81 @@
+## Possible solution for different reporting units for discards
+## requires a decision on size of catch on which to base mean weights
+## to include all data use 1
+## if a more representative sample is desired try 30 or 100?
+## the larger the sample size the smaller the resulting mean weight
+
+set_sample_size <- 1
+
+d <- data$catch
+
+## to focus on recent subset of catch data
+# d <- d$catch |>
+#   filter(year > 2005 & year < 2024) |>
+#   select(year, best_date, gear, fishery_sector, landed_kg, landed_pcs, discarded_kg, discarded_pcs)
+
+d[d$discarded_kg == 0 & d$discarded_pcs > 0, ]$discarded_kg <- NA
+d[d$discarded_pcs == 0 & d$discarded_kg > 0, ]$discarded_pcs <- NA
+
+# ## if we wanted to look at mean_weight by year
+# d <- d |> mutate(is_kg_na = ifelse(is.na(discarded_kg), "T", "F"),
+#                    is_pcs_na = ifelse(is.na(discarded_pcs), "T", "F"))
+#
+# d_sum <- d |> group_by(year, gear, is_kg_na, is_pcs_na) |> summarise(
+#   discarded_kg = sum(discarded_kg, na.rm = TRUE),
+#   discarded_pcs = sum(discarded_pcs, na.rm = TRUE),
+#   mean_weight = ifelse(is_kg_na == "F" & is_pcs_na == "F",
+#                        discarded_kg/discarded_pcs, NA),
+#   n = n()
+# ) |> distinct()
+#
+# d_sum |> filter(gear == "HOOK AND LINE") |> view()
+#
+# d_sum |>
+#   filter(!is.na(mean_weight)) |>
+#   select(year, gear, mean_weight, n) |>
+#   distinct() |>
+#   View()
+
+d2 <- d |>
+  group_by(fishery_sector, gear, is.na(discarded_kg), is.na(discarded_pcs)) |>
+  mutate(
+    mean_weight = mean(ifelse(
+      !is.na(discarded_kg) & !is.na(discarded_pcs) & discarded_pcs > set_sample_size,
+      discarded_kg/discarded_pcs, NA), na.rm = TRUE),
+    n = ifelse(
+      !is.na(discarded_kg) & !is.na(discarded_pcs) & discarded_pcs > set_sample_size,
+      discarded_pcs, NA),
+    mean_pcs_per_trip = mean(n, na.rm = TRUE),
+    n_w_both = ifelse(!is.na(discarded_kg) & !is.na(discarded_pcs), sum(!is.na(n)), NA)
+   ) |> ungroup() |>
+  group_by(fishery_sector, gear) |>
+  mutate(
+     mean_weight = mean(mean_weight, na.rm = TRUE),
+     mean_pcs_per_trip = mean(mean_pcs_per_trip, na.rm = TRUE),
+     n_w_both = mean(n_w_both, na.rm = TRUE),
+     n_w_kg = mean(ifelse(!is.na(discarded_kg) & discarded_kg > 0,
+                     sum(!is.na(discarded_kg)), NA), na.rm = TRUE),
+     n_w_pcs = mean(ifelse(!is.na(discarded_pcs) & discarded_pcs > 0,
+                      sum(!is.na(discarded_pcs)), NA), na.rm = TRUE),
+     n_trips = n()
+   ) |> ungroup() |>
+  mutate(
+    discarded_kg_obs = discarded_kg,
+    discarded_pcs_obs = discarded_pcs,
+    discarded_kg = ifelse(is.na(discarded_kg_obs),
+                          discarded_pcs_obs*mean_weight,
+                          discarded_kg_obs),
+    discarded_pcs = round(ifelse(is.na(discarded_pcs_obs),
+                          discarded_kg_obs/mean_weight,
+                          discarded_pcs_obs))
+  ) |>
+  relocate(
+    gear, year,
+    discarded_kg_obs, discarded_kg,
+    discarded_pcs_obs, discarded_pcs
+  )
+
+d2 |> select(gear, fishery_sector, mean_weight, mean_pcs_per_trip,
+             n_w_both, n_w_kg, n_w_pcs, n_trips) |>
+  distinct() |> View()
+