2021_w40_NBERpapers.Rmd

---
title: "2021_w40_NBERpapers"
author: "Albert Rapp"
date: "28 9 2021"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```


```{r}
library(tidyverse)
papers <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-09-28/papers.csv')
authors <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-09-28/authors.csv')
programs <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-09-28/programs.csv')
paper_authors <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-09-28/paper_authors.csv')
paper_programs <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-09-28/paper_programs.csv')
```

Let's bring these tables together by joining them.

```{r}
dat <- paper_programs %>% 
  left_join(programs, by = "program") %>% 
  left_join(paper_authors, by = "paper") %>% 
  left_join(authors, by = "author") %>% 
  left_join(papers, by = "paper") %>% 
  mutate(catalogue_group = str_sub(paper, 1, 1)) %>% 
  mutate(
    catalogue_group = case_when(
      catalogue_group == "h" ~ "Historical",
      catalogue_group == "t" ~ "Technical",
      catalogue_group == "w" ~ "General"
    )
  )
```


Let's try to manually do a micro vs. macro split violin plot over time.
First, take a look at the manual counts to see if that looks possible.

```{r}
counts <- dat %>% 
  filter(str_detect(program_category, "Micro") |  
           str_detect(program_category, "Macro")) %>% 
  count(program_category, year) %>% 
  mutate(
    n = if_else(str_detect(program_category, "Micro"), n, -n)
  ) 

count_plot <- counts %>% 
  ggplot(aes(x = n, y = factor(year), fill = program_category)) +
  geom_col() +
  scale_y_discrete(breaks = seq(1975, 2021, 5)) +
  coord_cartesian(xlim = c(-9000, 9000))
count_plot
```

This looks like it might be suitable for a decent plot.
Thus, compute the densities.

```{r}
estimate_density <- function(dat, variable) {
  dens <- dat %>% 
    filter(program_category == variable) %>% 
    pull(year) %>%
    abs() %>% 
    density()
  
  tibble(
    year = dens$x,
    count = dens$y,
    program_category = variable
  ) %>% 
  mutate(count = if_else(program_category != "Micro", -count, count))
}

  
dens <- dat %>% 
  estimate_density("Micro") %>% 
  bind_rows(dat %>% estimate_density("Macro/International")) 

uncropped <- dens %>% 
  ggplot(aes(count, year, fill = program_category)) +
  geom_polygon()
uncropped
```

Now crop to range

```{r}
year_range <- dat$year %>% range()

add_entries <- function(dens, variable) {
  dens %>% 
    filter(
      between(year, year_range[1], year_range[2]),
      program_category == variable
    ) %>% 
    bind_rows(
      tibble(
        year = year_range[1], 
        count =  0, 
        program_category = variable
      ), .) %>% 
    bind_rows(
      tibble(year = year_range[2], count =  0, program_category = variable)
    )
}

cropped_plot <- dens %>% 
  add_entries("Micro") %>% 
  bind_rows(dens %>% add_entries("Macro/International")) %>% 
  ggplot(aes(count, year, fill = program_category)) +
  geom_polygon() +
  coord_cartesian(xlim = c(-0.09, 0.09))
cropped_plot
```


Adjust scales and labels

```{r}
count_plot_adj <- count_plot +
  scale_fill_brewer(palette = "Set1") +
  scale_y_discrete(
    breaks = seq(1980, 2020, 10)
  ) +
  labs(
    x = element_blank(),
    y = "Year", 
    fill = "Program"
  )

croped_plot_adj <- cropped_plot +
  scale_fill_brewer(palette = "Set1") +
  scale_x_continuous(breaks = NULL) +
  scale_y_continuous(breaks = seq(1980, 2020, 10), minor_breaks = NULL) +
  labs(
    x = element_blank(),
    y = "Year", 
    fill = "Program"
  )

```


```{r}
library(patchwork)
arranged <- count_plot_adj + croped_plot_adj + 
  plot_annotation(
    title = "Macro vs. Micro",
    subtitle = "Comparing the Amount of Publications distributed by the NBER",
    caption = "A TidyTuesday contribution by @rappa753"
  ) +
  plot_layout(guides = "collect", heights = 1) &
  theme_light() +
  theme(
    legend.position = "top",
    text = element_text(size = 12, face = "bold")
  ) 
arranged
```

Save plot

```{r}
ggsave(
  "2021_w40_NBERpapers.png", 
  arranged,
  width = 30,
  height = 30 * 9 / 16,
  units = "cm"
)
```