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

knitr::opts_chunk$set(
  echo    = FALSE, 
  message = FALSE, 
  warning = FALSE
)

#library(remotes)
#remotes::install_github("ices-tools-dev/RDBEScore", ref = "dev")
library(RDBEScore)
library(DT)
library(dplyr)
library(plotly)
library(ggplot2)
library(leaflet)
library(formattable)
library(collapsibleTree)
library(scales)
#install.packages('devtools')
#devtools::install_github("ricardo-bion/ggtech", dependencies=TRUE)
library(ggtech)

# Suppress summarise info
#options(dplyr.summarise.inform = FALSE)

StrataEst <- readRDS("./testData/StrataEst.rds")
df <- StrataEst

df$Year = substr(df$stratumName, 0, 4)

df <- df  %>%
  filter_at(vars(est.total), any_vars((!is.na(.))))
  #%>% mutate(var.sd = sqrt(var.total)) 

# Design
DE <-df %>% filter(recType == 'DE') %>%
  rename('DE.Total' = est.total)


# Sampling Details
SD <-df %>% filter(recType == 'SD')%>%
  filter_at(vars(est.total), any_vars((!is.na(.)))) %>%
  rename('SD.Total' = est.total)

SD$Country <- substr(SD$Year, 0,2)

SD <- select(SD, -Year)

# vessel Selection
VS <-df %>% filter(recType == 'VS')


SD_DE <- full_join(SD,DE, by = c('parentTableStratum'='parentIDandStratum'))


#Set a few color variables to make our table more visually appealing

customGreen0 = "#DeF7E9"

customGreen = "#71CA97"

customRed = "#ff7f7f"

customRed0 = "#f9d1d1"

# scientific_10 <- function(x) {
#   parse(text=gsub("e", " %*% 10^", scales::scientific_format()(x)))
# }

Tree Diagram for standard error of Vessel Selection Strata

Plot showing the standard error for the vessel selection strata. The tree displays the parent strata, Design and Sampling Details. Click on a node to expand its child strata.

All_Data <- full_join(VS, 
                      SD_DE, 
                      by = c('parentTableStratum'='parentIDandStratum'))%>%
  select(stratumName.y, 
         stratumName.x, 
         stratumName,
         SD.Total, 
         DE.Total, 
         est.total,
         var.total,
         se.total,
         #var.sd, 
         Year.y, 
         Country)  %>%
  mutate(stratumName.y = 
           paste0(stratumName.y,': DE Total - ',
                  format(round(as.numeric(DE.Total), 0), big.mark=","))) %>%
  mutate(stratumName.x = 
           paste0(stratumName.x,': SD Total - ',
                  format(round(as.numeric(SD.Total), 0), big.mark=","))) %>%
  mutate(se.total = 
           paste0(stratumName,': VS stand. error - ',
                  format(round(as.numeric(se.total), 0), big.mark=","))) %>%
  rename('SD_Stratum' = stratumName.x,
         'Variance_VS' = se.total,
         'DE_Stratum' = stratumName.y)%>%
  arrange(Year.y, Country) %>%
  distinct()

All_Data$colour <- All_Data$var.total
levels(All_Data$colour) <- colorspace::rainbow_hcl(11)

p <- collapsibleTreeSummary(All_Data, 
                            c("Year.y",
                              "DE_Stratum",
                              "SD_Stratum",
                              "Variance_VS"),
                            nodeSize = "leafCount",
                            width = 800,
                            height = 600,
                            root = "All Data",
                            tooltip = FALSE,
                            collapsed = TRUE,
                            maxPercent = 70)
p

Data Visualisations {.tabset .tabset-pills}

Design

Graph showing the Estimated Total per Year, coloured by Design Stratum Name.

#Design Plot and Table


DE <- DE %>%
  select(recType, stratumName, DE.Total, se.total, Year, stratumName)
# # plot
DE_Plot <- ggplot(data = DE, aes(x = Year, y= DE.Total, fill = stratumName)) +
  geom_bar(stat='identity',  alpha= 0.7)+
   #scale_fill_brewer(palette = "Spectral") + 
  labs(title="Design Estimated Total by Year",x="Year", y = "Estimated Total") +
# scale_y_continuous(labels = scientific_10)+
 theme_tech(theme="airbnb")+
 theme(plot.title = element_text(color = "grey30"))+
  labs(fill = "DE Stratum Name")
ggplotly(DE_Plot)

formattable(DE,align =c("l","c","c","c","c", "c", "c", "c", "r"), 
      list(`Indicator Name` = 
      formatter("span", style = ~ style(color = "grey",font.weight = "bold")), 
            `DE.Total`= 
        color_tile(customGreen, customGreen0)
))

Sampling Details

Graph showing the Estimated Total by Stratum, coloured by Sampling Details Stratum Name.

#Sampling Details Plot and Table
SD <- SD %>%
  select(recType, stratumName, SD.Total, se.total, parentTableStratum) %>%
  mutate(parentTableStratum = sub('.', '', parentTableStratum)) 

SD_Plot <- ggplot(data = SD,
                  aes(x= parentTableStratum, y= SD.Total, fill = stratumName)) +
  geom_bar(stat = 'identity' ,  alpha= 0.7) +
  labs(title="Sampling Details Estimated Total by Stratum",
       x="DE Stratum", 
       y = "SD Estimated Total") +
  scale_fill_tech(theme="etsy") + 
  theme_tech(theme="airbnb")+
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) +
  theme(plot.title = element_text(color = "grey30"))+
  labs(fill = "SD Stratum Name")

ggplotly(SD_Plot)


formattable(SD, 
            align =c("c"), 
            list(`Indicator Name` = 
      formatter("span", style = ~ style(color = "grey",font.weight = "bold")), 
                 `SD.Total`= 
        color_tile(customGreen, customGreen0)
))

Vessel Selection

Graph 1 shows the Vessel Selection estimated total per Vessel Selection stratum name. The error bars show the standard error of the values with smaller bars depicting higher confidence levels. For the error bars with Standard Error value of NA, the confidence level is unknown.

Graph 2 displays the standard error by Vessel Selection stratum name coloured by Design stratum name.

Graph 3 displays the standard error by Vessel Selection stratum name coloured by the Sampling Details stratum name.

#### Vessel Selection Plot and Table ####

# estimated total with standard error bars

VS <- VS %>%
  mutate(error.min = est.total-se.total) %>%
  mutate(error.max = est.total+se.total)

VS1 <- left_join(VS, SD_DE, by = c('parentTableStratum'='parentIDandStratum'))

# error bar plot
VS_error <- ggplot(data=VS,
  aes(x=stratumName, y=est.total, text= paste("Std Error:", se.total))) +
  geom_errorbar(
    aes(x=stratumName, ymin=error.min, ymax=error.max), 
    width=0.2, colour="plum4", alpha=0.9, size=1.5) +
  geom_point(color="plum4", alpha= 0.7, size = 2) +
  labs(title="VS Estimated Total showing Standard Error",
       x="VS Stratum Name",
       y = "Estimated Total") +
  theme_tech(theme="airbnb") +
  theme(plot.title = element_text(color = "grey30")) +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))
ge<- ggplotly(VS_error)
ge$x$data[[1]]$hoverinfo <- "none"

ge

# Standard error for VS by VS stratum name coloured by DE Stratum Name
VS_scatter1 <-ggplot(data=VS1,
                      aes(x=stratumName, y=se.total, fill = stratumName.y)) +
  geom_point(color="cadetblue", alpha= 0.7, size=4) +
  labs(title="VS Stand. Error by VS Stratum Name",
       x="VS Stratum Name",
       y = "Standard error") +
  theme_tech(theme="airbnb") +
  theme(plot.title = element_text(color = "grey30")) +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) +
  labs(fill = "DE Stratum Name")
ggplotly(VS_scatter1)

# Standard error for VS by VS stratum name coloured by SD Stratum Name
VS_scatter2 <-ggplot(data=VS1,
                     aes(x=stratumName, y=se.total, fill = stratumName.x)) +
  geom_point(color="skyblue4", alpha= 0.7, size=4) +
  scale_fill_tech(theme="airbnb") +
  labs(title="VS Stand. Error by SD Stratum Name",
       x="VS Stratum Name",
       y = "Standard error") +
  theme_tech(theme="airbnb") +
  theme(plot.title = element_text(color = "grey30")) +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) +
  labs(fill = "SD Stratum Name")
ggplotly(VS_scatter2)

# tidy df for table
VS2 <- VS %>%  
  select(recType, stratumName, parentTableStratum, est.total, se.total)
  #select(recType, stratumName, parentTableStratum, est.total, var.total)%>%
  #filter(!is.na(var.total))




# table
formattable(VS2, align = "c", 
            list(`Indicator Name` = 
      formatter("span", style = ~ style(color = "grey",font.weight = "bold")), 
                 `est.total` = 
      color_tile(customGreen, customGreen0),
                  `se.total` = 
        color_tile(customRed, customRed0)
))

Fishing Trip

This graphs shows the Fishing Trip estimated totals by Vessel Selection stratum name.

# Fishing Trip

FT <- df %>% filter(recType == 'FT') %>%
  mutate(parentTableStratum = sub('.*:', '', parentTableStratum)) 


FT_scatter <-ggplot(data=FT, aes(x=parentTableStratum, y=est.total)) +
  geom_point(color="seagreen", alpha= 0.7, size=4) +
  labs(title="FT Estimated Total by VS Stratum",
       x = "VS Stratum",
       y = "Estimated Total") +
 theme_tech(theme="airbnb") +
 theme(plot.title = element_text(color = "grey30")) +
    theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))
ggplotly(FT_scatter)

Fishing Operation

Graph 1 shows the Fishing Operation estimated total per Vessel Selection stratum name. The error bars show the standard error of the values with smaller bars depicting higher confidence levels. For the error bars with Standard Error value of NA, the confidence level is unknown.

Graph 2 displays the standard error by Vessel Selection stratum name.

# Fishing Operation

FO <- df %>% filter(recType == 'FO') 

FO <- left_join(FO, FT, by = c('parentTableStratum'='parentIDandStratum'))

FO <- FO %>%
  mutate(error.min = est.total.x-se.total.x) %>%
  mutate(error.max = est.total.x+se.total.x) %>%
  rename("FO.est.total"=est.total.x)%>%
  rename("VS.stratumName"= parentTableStratum.y)

#FO <- FO %>% filter(!is.na(se.total.x))

FO_error <- ggplot(data=FO,
              aes(x=VS.stratumName, 
                  y=FO.est.total, 
                  text= paste("Std Error:", se.total.x))) +
  geom_errorbar(
      aes(x=VS.stratumName, ymin=error.min, ymax=error.max)
          ,colour="maroon", width=0.2,  alpha=0.9, size=1.5) +
  geom_point(color="maroon", alpha= 0.7, size = 2) +
  labs(title="FO Estimated Total showing Standard Error",
       x="VS Stratum Name",
       y = "Estimated Total") +
  theme_tech(theme="airbnb") +
  theme(plot.title = element_text(color = "grey30")) +
  theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))
gg<- ggplotly(FO_error)
gg$x$data[[1]]$hoverinfo <- "none"

gg

FO_scatter1 <-ggplot(data=FO, aes(x=VS.stratumName, y=se.total.x))+
  geom_point(color="aquamarine4", alpha= 0.7, size = 4) +
  labs(title="FO Stand. Error by VS Stratum",
       x="VS Stratum Name",
       y = "Standard error") +
 theme_tech(theme="airbnb") +
 theme(plot.title = element_text(color = "grey30")) +
      theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) 
ggplotly(FO_scatter1)

# FO_scatter2 <-ggplot(data=FO, aes(x=parentTableStratum.y, y=est.total.x))+
#   geom_point(color="sandybrown", alpha= 0.7, size = 4) +
#   labs(title="FO Estimated Total by VS Stratum",x="VS Stratum Name", y = "Estimated Total")+
#  theme_tech(theme="airbnb")+
#  theme(plot.title = element_text(color = "grey30"))+
#       theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1))
# ggplotly(FO_scatter2)

Species Selection

This graphs shows the Species Selection estimated totals by Vessel Selection stratum name.

# Species Selection plot (no standard error available)

SS <- df %>% filter(recType == 'SS') 

SS1 <- left_join(SS, FO, by = c('parentTableStratum'='parentIDandStratum'))%>%
  select(VS.stratumName, est.total,parentIDandStratum, parentTableStratum)


SS_scatter <-ggplot(data=SS1, aes(x=VS.stratumName, y=est.total)) +
  geom_point(color="cadetblue", alpha= 0.7, size = 4) +
  labs(title="SS Estimated Total by VS Stratum",
       x="VS Stratum",
       y = "Estimated Total") +
 theme_tech(theme="airbnb") +
 theme(plot.title = element_text(color = "grey30")) +
 theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust=1)) 
ggplotly(SS_scatter)

Sample

This graphs shows the Sample estimated totals by Vessel Selection stratum name.

# Sampling plot (no standard error available)

SA <- df %>% filter(recType == 'SA')
SA <- left_join(SA, SS1, by = c('parentTableStratum'='parentIDandStratum'))


SA_scatter <-ggplot(data=SA, aes(x=VS.stratumName, y=est.total.x))+
  geom_point(color="deepskyblue3", alpha= 0.7, size = 4) +
  labs(title="SA Estimated Total by VS Stratum",
       x="VS Stratum", 
       y = "Estimated Total") +
 theme_tech(theme="airbnb") +
 theme(plot.title = element_text(color = "grey30")) +
      theme(axis.text.x = element_blank()) 
ggplotly(SA_scatter)

3D Plot

3D graph showing the standard error and estimated total per Year, coloured by Sampling Details stratum name.

#3D Plot

joinedUpper <- left_join(VS, 
                         SD_DE, 
                         by = c('parentTableStratum'='parentIDandStratum')) %>%
  select(DE.Total, 
         est.total,
         var.total,
         #var.sd, 
         se.total,
         Year.y, 
         stratumName.x, 
         se.total) %>%
  rename('VS.Total' = est.total,
         #'VS.sd' = var.sd,
         'VS.se' = se.total,
         'VS.Var.Total' = var.total)


p <- plot_ly(joinedUpper, 
             x = ~as.factor(Year.y), 
             #y = ~VS.sd,  
             y = ~VS.se, 
             z= ~VS.Total, color = ~stratumName.x)
p <- p %>% add_markers()
p <- p %>% 
  layout(title = paste0('Stand. Error and Estimated Total for',
                          'Vessel Selection per Year by SD Stratum'),
         scene = list(xaxis = list(tickvals= ~Year.y, title = 'Year'),
         yaxis = list(title = 'Stand. Error VS'),
         zaxis = list(title = 'Estimated Total VS')))

p

# Histograms

# p<-ggplot(joinedUpper, aes(x=DE.Total)) + 
#   geom_histogram(color="ivory3", fill="lightblue4")+ 
#   geom_vline(aes(xintercept=mean(DE.Total)),
#             color="red4", linetype="dashed", size=1) +
#     labs(title=" Details Estimated Total",x="Estimated Total", y = "Count")+
#  theme_tech(theme="airbnb")+
#  theme(plot.title = element_text(color = "grey30"))
# ggplotly(p)
# 
# p2<-ggplot(joinedUpper, aes(x=VS.Total)) + 
#   geom_histogram(color="ivory4", fill="lightskyblue3")+ 
#   geom_vline(aes(xintercept=mean(VS.Total)),
#             color="red4", linetype="dashed", size=1)+
#     labs(title="Vessel Selection Estimated Total",x="Estimated Total", y = "Count")+
#  theme_tech(theme="airbnb")+
#  theme(plot.title = element_text(color = "grey30"))
# ggplotly(p2)

# p3<-ggplot(joinedUpper, aes(x=VS.Var.Total)) + 
#  geom_histogram(aes(y=..density..), colour="ivory4", fill="lightcyan3")+
#  geom_density(alpha=.5, fill="red4") +
#     labs(title="Vessel Selection Variance Density of Values",x="Variance", y = "Density")+
#  theme_tech(theme="airbnb")+
#  theme(plot.title = element_text(color = "grey30"))+
#     scale_x_continuous(labels = label_number_si(accuracy=0.1))
# ggplotly(p3)