Déterminer la taille des échantillons par site

Author

Steve Vissault

Published

January 5, 2024

On compte le nombre de mesures pour chaque combinaison.

source("src/itgr_measurements.R")
contaminants <- itgr_measurements()

data_count <- contaminants |> 
    dplyr::mutate(Year = as.integer(Year)) |>
    toxbox::uncensored(cols = "value", keep_cens = TRUE) |>
    dplyr::group_by(Location, Year, Species, conpound_family, variable) |>
    dplyr::summarise(n = dplyr::n(), n_censored = sum(value_censored))

Warning: There was 1 warning in `dplyr::mutate()`.
ℹ In argument: `dplyr::across(tidyselect::all_of(cols), remove_cens)`.
Caused by warning:
! NAs introduced by coercion

`summarise()` has grouped output by 'Location', 'Year', 'Species',
'conpound_family'. You can override using the `.groups` argument.

data_count |>
    reactable::reactable()

Criteria 1. On retire les sites pour lesquelles, il y a une seule année de collecte (impossibilité de calibrer un modèle de tendance temporelle).

Voici la liste des sites pour lesquelles, on a une seule année de mesure

crit1 <- contaminants |> 
    dplyr::select(Location, Year) |>
    dplyr::distinct() |>
    dplyr::group_by(Location) |>
    dplyr::count() |>
    dplyr::filter(n == 1) |>
    dplyr::pull(Location)

cat(paste("*", crit1), sep="\n")

Archipel de Mingan (approximative location)
Beauharnois
Betchouanes
Bonaventure
Deslauriers
Escuminac
Gros Pot
Ile Beauséjour
Ile Bellechasse
Ile Bienville
Ile Blanche
Ile Dodens
Ile Dowker
Ile Nue
Ile Saint-Barnabé
Ile a Bouleau de Terre
Ile aux Pommes
Ile de la Madeleine
Ile de la Mine
Ile du Petit Caoui
Ile du Petit Caouis
Ile à Calculot des Betchouanes
Iles aux Loups Marins
Iles de la Madeleine
Ilets Boisés
Ilets De Quen
Kamouraska
Kukuministuk (Caye à Cochons)
Lac Bleu
Lac Evans
Lac Saint-Bernard
Lac St-Pierre
Lac à la Carpe
Long Island
Miller island
Missipinuk (Grosse Romaine)
Petit Lac Jacques Cartier
Petite Ile Ste-Geneviève
Pointe Saint-Pancrace
Prairie du Castor
Riviere George

On retire ces sites pour l’étude de la taille de l’échantillon

data_count <- data_count |>
    dplyr::filter(!(Location %in% crit1)) |>
    dplyr::ungroup()

Site Battures aux Loups Marins

Site Bonaventure Island

Site Grande Ile

Site Ile Carillon

Site Ile Dickerson

Site Ile Eaton

Site Ile Laval

Site Ile Manowin

Site Ile Matane

Site Ile Saint-Bernard

Site Ile Steamboat

Site Ile Villemomble

Site Ile aux Basques

Site Ile aux Herons

Site Ile aux Hérons

Site Ile aux Oeufs

Site Ile de Carillon

Site Ile de la Corneille

Site Ile du Bic

Site Ile du Corossol

Site Iles Sainte-Marie

Site Maria

Site Petit lac Jacques-Cartier

Site Refuge Watshishou

Site Yamachiche

--- title: "Déterminer la taille des échantillons par site" author: "Steve Vissault" date: "2024-01-05" --- On compte le nombre de mesures pour chaque combinaison. ```{R, results = "asis"} source("src/itgr_measurements.R") contaminants <- itgr_measurements() data_count <- contaminants |> dplyr::mutate(Year = as.integer(Year)) |> toxbox::uncensored(cols = "value", keep_cens = TRUE) |> dplyr::group_by(Location, Year, Species, conpound_family, variable) |> dplyr::summarise(n = dplyr::n(), n_censored = sum(value_censored)) ``` ```{r} data_count |> reactable::reactable() ``` ### Criteria 1. On retire les sites pour lesquelles, il y a une seule année de collecte (impossibilité de calibrer un modèle de tendance temporelle). Voici la liste des sites pour lesquelles, on a une seule année de mesure ```{R, results = "asis"} crit1 <- contaminants |> dplyr::select(Location, Year) |> dplyr::distinct() |> dplyr::group_by(Location) |> dplyr::count() |> dplyr::filter(n == 1) |> dplyr::pull(Location) cat(paste("*", crit1), sep="\n") ``` On retire ces sites pour l'étude de la taille de l'échantillon ```{r} data_count <- data_count |> dplyr::filter(!(Location %in% crit1)) |> dplyr::ungroup() ``` ```{r, include = FALSE, eval = FALSE} library(ggplot2) data_count <- data_count |> dplyr::mutate(n = cut(n, breaks = c(1, 2, 3, 5, 10, Inf), labels = c("1", "2", "3", "5-9", ">=10"), right = FALSE)) |> dplyr::mutate(n_censored = cut(n_censored, breaks = c(1, 2, 3, 5, 10, Inf), labels = c("1", "2", "3", "5-9", ">=10"), right = FALSE)) sites <- data_count |> dplyr::pull(Location) |> unique() for(s in sites){ conpounds <- data_count |> dplyr::filter(Location == s) |> dplyr::pull(conpound_family) |> unique() figs <- list() for (c in seq_along(conpounds)) { d <- data_count |> dplyr::filter(conpound_family == conpounds[c] & Location == s) |> tidyr::complete( tidyr::nesting(Location, Species, conpound_family, variable), Year = min(data_count$Year):max(data_count$Year), fill = list(n = NA) ) figs[[c]] <- ggplot(d, aes(x = Year, y = variable, fill = n)) + geom_tile() + # coord_fixed(xlim = range(data_count$Year)) + ggtitle(conpounds[c]) + scale_x_continuous(breaks = seq(min(data_count$Year), max(data_count$Year), 1), limits = range(data_count$Year)) + viridis::scale_fill_viridis(discrete = TRUE, na.translate = FALSE, drop = FALSE) + theme_grey(base_size = 20) + theme( axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1), panel.grid.major = element_blank(), panel.grid.minor = element_blank() ) + xlab("Year of measurement") + ylab("Contaminant") + labs(fill='Measurements (n)') + facet_wrap(~Species, ncol = 1) } # rename figs list names(figs) <- conpounds # Calcul les tailles des figures tmp <- data_count |> dplyr::filter(Location == s) |> dplyr::group_by(Location, conpound_family) |> dplyr::summarize(h = dplyr::n_distinct(variable)) |> dplyr::mutate(tot = sum(h) + 10) |> dplyr::mutate(h = h / sum(h)) |> dplyr::arrange(factor(conpound_family, levels = names(figs))) heights <- tmp |> dplyr::pull(h) tot <- tmp |> dplyr::pull(tot) |> unique() filename <- paste0(janitor::make_clean_names(s), ".png") p <- patchwork::wrap_plots(figs, ncol = 1, heights = heights) ggsave( p, file = file.path("figures", filename), create.dir = TRUE, dpi = 300, width = 50, height = tot, units = "cm", limitsize = FALSE ) } ``` ```{R, echo = FALSE, results = "asis"} sites <- data_count |> dplyr::pull(Location) |> unique() for (s in sites){ cat("\n### Site ", s, "\n") cat(paste0("![](", file.path("figures", paste0(janitor::make_clean_names(s), ".png")), ")"), "\n") cat("\n") } ```