In ices-tools-dev/icesRDBES: Functions for the ICES Regional Database and Estimation System (RDBES)

knitr::opts_chunk$set(echo = TRUE)

Example of functions to highlight sources of potential bias

# Load our functions
library(dplyr)
library(RDBEScore)
source("summariseSelectionMethods.R")
library(kableExtra)

# IMPORTANT: Hack to stop write.csv changing numbers to scientific notation
options(scipen=500) # big number of digits

## STEP 1) LOAD OUR DATA

myRDBESData <- createRDBESDataObject("..\\..\\..\\tests\\testthat\\h1_v_1_19")

Selection methods

The first function highlights non-probabilistic selection methods, and the number of units that were intended to be sampled but weren't - these are both potentially sources of bias.

myResult <- summariseSelectionMethods(objectToSummarise = myRDBESData,
                                      yearToUse = 1965,
                                      country = 'ZW',
                                      hierarchyToSummarise = 'H1')


for (myTable in names(myResult$selectionMethodSummary)){

# This is the data we'll print - after soem small formatting changes
myDataToPrint <- myResult$selectionMethodSummary[[myTable]]
# change NA to blanks
myDataToPrint[is.na(myDataToPrint)] <- ""
# get the "non caculated" column numbers - we'll format these differently
myNonCalcColumns <- which(names(myDataToPrint) %in% names(myDataToPrint)[grepl('^..numberNotSampledCalc$',names(myDataToPrint))])


# Print the selection methid summary
p <- myDataToPrint %>% 
    mutate_all(~cell_spec(
        .x,
        background = ifelse(.x %in% c('NPCLQS-O', 'NPCLQS-T', 'NPCS', 'NPJS', 'NPQSRSWOR', 'NPQSRSWR', 'NPQSYSS'), 'red','white'),
        color = ifelse(.x %in% c('NPCLQS-O', 'NPCLQS-T', 'NPCS', 'NPJS', 'NPQSRSWOR', 'NPQSRSWR', 'NPQSYSS'), 'white','black'))) %>%
  kbl(escape = F, caption = myTable) %>%
  column_spec(myNonCalcColumns, bold=T) %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"))

print(p)

}


# This is the data we'll print - after soem small formatting changes
dataToPrint <- myResult$summaryDataJoined
# change NA to blanks
dataToPrint[is.na(dataToPrint)] <- ""
# get the "non caculated" column numbers - we'll format these differently
myNonCalcColumns <- which(names(dataToPrint) %in% names(dataToPrint)[grepl('^..numberNotSampledCalc$',names(dataToPrint))])

# Print the more detailed infromation by stratum
dataToPrint %>% 
    mutate_all(~cell_spec(
        .x,
        background = ifelse(.x %in% c('NPCLQS-O', 'NPCLQS-T', 'NPCS', 'NPJS', 'NPQSRSWOR', 'NPQSRSWR', 'NPQSYSS'), 'red','white'),
        color = ifelse(.x %in% c('NPCLQS-O', 'NPCLQS-T', 'NPCS', 'NPJS', 'NPQSRSWOR', 'NPQSRSWR', 'NPQSYSS'), 'white','black'))) %>%
  kbl(escape = F) %>%
  column_spec(myNonCalcColumns, bold=T) %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed", "responsive"))

ices-tools-dev/icesRDBES documentation built on April 17, 2025, 1:58 p.m.