2021_w42_global_fishing.Rmd

---
title: "2021_w42_global_fishing"
output: html_document
editor_options: 
  chunk_output_type: console
---
```{r}
library(tidyverse)
farmed <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-10-12/aquaculture-farmed-fish-production.csv')
captured_vs_farmed <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-10-12/capture-fisheries-vs-aquaculture.csv')
captured <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-10-12/capture-fishery-production.csv')
consumption <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-10-12/fish-and-seafood-consumption-per-capita.csv')
stock <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-10-12/fish-stocks-within-sustainable-levels.csv')
fishery <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-10-12/global-fishery-catch-by-sector.csv')
production <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-10-12/seafood-and-fish-production-thousand-tonnes.csv')

```


```{r}
fishery_longer <- fishery %>% 
  pivot_longer(-(1:3), names_to = "sector", values_to = "amount")
```

```{r}
fishery_longer %>% count(Entity)
global_fishery_sectors <- fishery_longer %>% janitor::clean_names()
```

So, fishery contains only entity "World".
Thus, let"s name that `global_fishery_sectors`

```{r}
amounts_plot <- global_fishery_sectors %>% 
  filter(sector != "Recreational") %>% 
  mutate(
    sector = fct_reorder(sector, amount)
  ) %>% 
  ggplot(aes(year, amount, fill = sector)) +
  geom_col() +
  # geom_label(
  #   data = filter(global_fishery_sectors, sector != "Recreational", year == 2010),
  #   aes(x = year, y = amount, label = sector),
  #   position = position_stack(vjust = 0.5),
  #   color = 'white',
  #   size = 8
  # ) +
  scale_y_continuous(labels = scales::number) +
  scale_fill_brewer(palette = "Set1") +
  theme_minimal() +
  theme(
    legend.position = c(0.25, 0.8),
    text = element_text(face = 'bold', size = 12),
    plot.title.position = 'plot'
  ) +
  labs(
    x = "Year",
    y = "Tonnes of fish",
    fill = "Sector",
    subtitle = "Amount of world-wide fish production"
  )
amounts_plot
```


Take a look at `stock` and rearrange the columns a bit.

```{r}
sustainable_overexploited_shares <- stock %>%
  janitor::clean_names() %>% 
  select(-code) %>% 
  rename_with(~case_when(
    str_detect(., "biologically") ~ "sustainable",
    str_detect(., "overexploited") ~ "overexploited",
    TRUE ~ .
  )) %>% 
  pivot_longer(3:4, names_to = "type", values_to = "share")
```


```{r}
sustainable_overexploited_shares %>% 
  count(entity)
```


Since there are only four observations in the regions other than "World", let's ignore them.

```{r}
sustainable_overexploited_shares %>% 
  filter(entity == "World") %>% 
  ggplot(aes(x = year, y = share, fill = type)) +
  geom_col(position = 'fill')
```

There are a lot of missing years. 
Ideally, I would like to fill the gaps, by letting the values stay constant.
Also, I want to indicate by some shading that these values are only "guesses".
Overall, I hope that this makes a visual comparison over time easier for the reader.

```{r}
filled_shares <- sustainable_overexploited_shares %>% 
  filter(entity == "World") %>% 
  complete(year = 1974:2017, type = c('sustainable', 'overexploited')) %>% 
  mutate(filled = if_else(is.na(share), T, F)) %>% 
  group_by(type) %>% 
  group_split() %>% 
  map(~fill(., entity, type, share)) %>% 
  bind_rows()

shares_plot <- filled_shares %>% 
  mutate(fill_dummy = str_c(type, filled, sep = '_')) %>% 
  ggplot(aes(x = year, y = share, fill = type, alpha = filled)) +
  geom_col(position = 'fill') +
  geom_label(
    x = 1985, 
    y = 0.10, 
    label = 'sustainable',
    fontface = "bold",
    size = 12,
    color = 'white'
    ) +
  geom_label(
    x = 2005, 
    y = 0.90, 
    label = 'overexploited',
    fontface = "bold",
    size = 12,
    color = 'white',
    fill = RColorBrewer::brewer.pal(3, "Set1")[1]
    ) +
  scale_alpha_manual(values = c(1, 0.7)) +
  scale_fill_brewer(palette = "Set1") +
  theme_minimal() +
  scale_y_continuous(labels = scales::percent) +
  labs(
    x = "Year", 
    y = element_blank(), 
    subtitle = "Share of sustainable World-wide fishing",
    caption = "Lighter tones indicate missing values being filled with last available data"
  ) +
  theme(
    legend.position = 'none',
    plot.title.position = 'plot',
    text = element_text(size = 12, face = 'bold')
  )
shares_plot
```

Arrange both plots

```{r}
library(patchwork)
collected_plot <- amounts_plot + shares_plot +
  plot_annotation(
    title = "Even though the overall fish-production decreased in recent years, the share of overexploited fish stock increased"
  ) &
  theme(text = element_text(size = 14, face = 'bold'))
collected_plot
ggsave(
  '2021_w42_global_fishing.png',
  collected_plot,
  width = 40,
  height = 40 * 9 / 16, 
  units = 'cm'
)
```