In eriqande/lowergranite: assessing variability of abundances of salmon in different strata

Note that to get this to really be a vignette, I must insert this at the top:

<!--
%\VignetteEngine{knitr::knitr}
%\VignetteIndexEntry{quick overview of use/flow of functions}
-->

Introduction

I am just using this as a place to store some notes and write down some examples of the use of some of the functions. To start off you need the ''lowergranite'' library:

library(lowergranite)

Dealing with rare strata

The joint CI calculations take a beating when there are groups of fish that occur so infrequently in the data that there are iterations of the bootstrapping in which those groups never occur in any year. Kirk and Mike seemed to suggest that you may as well just drop those groups from the analysis. To faciliate that I have added a drop.these.groups option to run_boot_gsi_analysis(). You just pass it a character vector of the stocks-by-age or stock-by-sex columns you want to drop (the naming has to follow the "double-period" convention of the naming of columns in the data frames).

Let's see how this works. First, we run an analysis without dropping anyone. We don't do a long run here because I don't want to take too long making this document:

no_drop <- run_boot_gsi_analysis(lg_test_data$SH11.Stock.Age.With.Periods, stock_group_start_col=9, B=100, nsim=50)

And see what the output looks like, at least for the first four pops

no_drop[,1:4]

OK, now, let's do the same thing, but drop the pops that we want to. Note that I have those pops to drop for steelhead (the ones Kirk dropped in the analysis he emailed around on April 14) in the package data:

lg_test_data$sh.droppers

So, we run it:

with_drop <- run_boot_gsi_analysis(lg_test_data$SH11.Stock.Age.With.Periods, stock_group_start_col=9, B=100, nsim=50, drop.these.groups=lg_test_data$sh.droppers)

And the results we get (first five pops) are:

with_drop[, 1:4]

Checking that it works with the gsi_sim option

Let's give this a whirl and see if it works with gsi assignments:

# note that this fails...probably because there is a 0 length vector somewhere 
with_gsi_no_drop <- run_boot_gsi_analysis(lg_test_data$SH11.Stock.Age.With.Periods, stock_group_start_col=9, B=100, nsim=50, DO_GSI_ON_PROP=T)

eriqande/lowergranite documentation built on May 16, 2019, 8:47 a.m.