3.1_caa_comparison.Rmd

---
title: "Catch-at-age: comparison"
date: "`r Sys.Date()`"
output:
  html_document:
    toc: true
    toc_depth: 3
    toc_float:
      collapsed: FALSE
    code_folding: show
    number_sections: TRUE
---

# SET-UP    
```{r markdown, echo=F}
library(knitr)
opts_chunk$set(echo = T, collapse = T, fig.align = "center", fig.width = 9, fig.height = 6)
options(width = 140)
```    

## settings

```{r settings, message = F}
source('0.0_settings.R')
```

## load all data

```{r data, message = F}
# New caa
load(paste0(dir.rdat, "caa.Rdata"))

# Old caa and waa
year <- 2018
repo <- "https://github.com/iml-assess/mackerel_assessment/blob/master/"
caa.old <- read.ices(url(paste0(repo,'data/',year,'/cn.dat',"?raw=true")))
waa.old <- read.ices(url(paste0(repo,'data/',year,'/cw.dat',"?raw=true")))
```

## Bindind and scaling
```{r wrangle, message = F}
caa.old <- reshape2::melt(as.matrix(caa.old),varnames=c('year','age'),value.name='caan.old')
waa.old <- reshape2::melt(as.matrix(waa.old),varnames=c('year','age'),value.name='waa.old')

caa.comp <- Reduce(function(x, y) merge(x, y, all.x=TRUE), list(caa, caa.old, waa.old))
caa.comp <- caa.comp[caa.comp$year %in% unique(caa.old$year),]                                                      # remove the recent years
caa.comp <- ddply(caa.comp,c('year'),mutate,caan.rel=caan/max(caan),caan.old.rel=caan.old/max(caan.old))            # compare proportions instead of absolutes
```

# CAA PLOTS {.tabset}
 
## Side-wise absolute
```{r plot 1}
grid.arrange(
    ggplot(caa.comp,aes(x=year,y=age,size=caan))+geom_point(alpha=0.8)+ scale_size(range = c(1,8))+labs(title='NEW (absolute)'),
    ggplot(caa.comp,aes(x=year,y=age,size=caan.old))+geom_point(alpha=0.8)+ scale_size(range = c(1,8))+labs(title='OLD (absolute)')
)
```

## Side-wise relative
```{r plot2}
grid.arrange(
    ggplot(caa.comp,aes(x=year,y=age,size=caan.rel))+geom_point(alpha=0.8)+ scale_size(range = c(1,5))+labs(title='NEW (rel by year)'),
    ggplot(caa.comp,aes(x=year,y=age,size=caan.old.rel))+geom_point(alpha=0.8)+ scale_size(range = c(1,5))+labs(title='OLD (rel by year)')
)
```

## Residual absolute
If the difference is more than 100% the dots/triangles are black
```{r plot 3}
caa.comp$res <- with(caa.comp,(caan-caan.old)/caan.old)
caa.comp$big <- ifelse(abs(caa.comp$res)>100,TRUE,FALSE)
caa.comp$dir <- ifelse(caa.comp$res<0,'-',ifelse(caa.comp$res==0,'0','+'))
ggplot()+
    geom_point(data=caa.comp[caa.comp$big==FALSE,],aes(x=year,y=age,col=abs(res),shape=dir),size=4)+
    geom_point(data=caa.comp[caa.comp$big==TRUE,],aes(x=year,y=age,shape=dir),col='black',size=4)+
    labs(title='(NEW-OLD)/OLD')+
    scale_color_viridis_c(direction = -1)
```

## Residual relative
```{r plot 4}
caa.comp$res <- with(caa.comp,caan.rel-caan.old.rel)
caa.comp$col <- ifelse(caa.comp$res<0,'-',ifelse(caa.comp$res==0,'0','+'))
ggplot(caa.comp,aes(x=year,y=age,size=abs(res),col=col))+
    geom_point(alpha=0.8)+
    scale_color_manual(values=c('darkred','darkgreen','black'))+
    labs(title='NEW (rel) - OLD (rel)')
```

## Correlation
```{r plot 5}
corr <- ddply(caa.comp[!is.na(caa.comp$caan.rel) & !is.na(caa.comp$caan.old.rel),],c('age'),summarise,corr=cor(caan.rel,caan.old.rel))
ggplot(caa.comp,aes(x=caan.rel,y=caan.old.rel))+
    geom_point()+
    geom_text(aes(label=year),size=2,hjust=0,vjust=0)+
    geom_text(data=corr,aes(x=-Inf,y=Inf,label=round(corr,2)),vjust=1,hjust=0,col='blue')+
    geom_abline(slope=1,intercept=0)+
    facet_wrap(~age,scale='free')
```

# WAA PLOTS{.tabset}

## side-wise
```{r plot 1b}
grid.arrange(
    ggplot(caa.comp,aes(x=year,y=waa,group=age))+geom_line(aes(color=age),size=1)+scale_color_viridis()+labs(title='NEW'),
    ggplot(caa.comp,aes(x=year,y=waa.old,group=age))+geom_line(aes(color=age),size=1)+scale_color_viridis()+labs(title='OLD')
)
```

## side-wise by age
```{r plot 2b}
df <- reshape2::melt(caa.comp[,c('year','age','waa','waa.old')],id=c('year','age'))
ggplot(df,aes(x=year,y=value,group=variable,col=variable))+
  geom_line(size=1,alpha=0.6)+
  facet_wrap(~age,scale='free')+
  scale_color_manual(values=c('red','black'))
```

## Correlation
```{r plot 3b}
corr <- ddply(caa.comp[!is.na(caa.comp$waa.old) & !is.na(caa.comp$waa),],c('age'),summarise,corr=cor(waa,waa.old))
ggplot(caa.comp,aes(x=waa,y=waa.old))+
    geom_point()+
    geom_text(aes(label=year),size=2,hjust=0,vjust=0)+
    geom_text(data=corr,aes(x=-Inf,y=Inf,label=round(corr,2)),vjust=1,hjust=0,col='blue')+
    geom_abline(slope=1,intercept=0)+
    facet_wrap(~age,scale='free')
```