us_phds.Rmd

---
title: "US PhDs"
output: html_document
jupyter:
  jupytext:
    text_representation:
      extension: .Rmd
      format_name: rmarkdown
      format_version: '1.2'
      jupytext_version: 1.3.2
  kernelspec:
    display_name: Python 3
    language: python
    name: python3
---

```{python pytabs={'class': 'pytabs-1', 'name': 'R'}}
import rpy2

# %load_ext rpy2.ipython

from qgrid import show_grid

```

## Preprocessing

```{r pytabs={'class': 'pytabs-2', 'name': 'R'}}

library(tidyverse)
theme_set(theme_light())

# Major field of study
major_fields_raw <- readxl::read_xlsx("./data/sed17-sr-tab012.xlsx",
                  skip = 3)
```

```{r pytabs={'class': 'pytabs-3', 'name': 'R'}}

major_fields <- major_fields_raw %>%
  rename(field = `Field of study`) %>%
  gather(key, value, -field) %>%
  mutate(year = as.numeric(ifelse(str_detect(key, "X__"), NA, key)),
         type = ifelse(!str_detect(value, "Number|Percent"), NA, value),
         value = as.numeric(value)) %>%
  fill(year, type) %>%
  select(-key) %>%
  filter(!is.na(value)) %>%
  spread(type, value)

# major_fields_raw
```

## Plotting degrees over time for 6 degrees

```{r pytabs={'class': 'pytabs-4', 'name': 'R'}}

fine_fields <- readxl::read_xlsx("./data/sed17-sr-tab013.xlsx",
                  skip = 3) %>%
  rename(field = 1) %>%
  gather(year, number, -field) %>%
  mutate(year = as.numeric(year),
         number = as.numeric(number)) %>%
  filter(!is.na(number))
```

```{r pytabs={'class': 'pytabs-5', 'name': 'R'}}

set.seed(42)
field_subset = sample(unique(fine_fields$field), 6)

fine_fields %>%
  filter(field %in% field_subset) %>%
  ggplot(aes(year, number, color = field)) +
  geom_line()
```

```{r pytabs={'class': 'pytabs-6', 'name': 'R'}}

# get the broad field names and the major field names

sex <- c("All", "Male", "Female", "All doctorate recipientsa", "All fieldsa")

broad_fields <- readxl::read_xlsx("./data/sed17-sr-tab014.xlsx", skip = 4) %>%
  rename(field = 1) %>%
  filter(!field %in% sex) %>%
  mutate(field = fct_recode(field,
                            "Life sciences" = "Life sciencesb",
                            "Other" = "Otherc")) %>%
  pull(field) %>%
  as.character()
```


```{r pytabs={'class': 'pytabs-7', 'name': 'R'}}

recipients_year_field_sex <- readxl::read_xlsx("./data/sed17-sr-tab015.xlsx", skip = 3) %>%
  rename(field = 1) %>%
  select(-contains("change")) %>%
  mutate(field = as.character(fct_recode(field, "All" = "All doctorate recipientsa",
                                         "Other" = "Otherb")),
         sex = if_else(field %in% sex, field, NA_character_),
         broad_field = ifelse(field %in% broad_fields, field, NA)) %>%
  fill(sex, broad_field) %>%
  gather(year, number, -sex, -broad_field, -field) %>%
  mutate(year = as.numeric(year),
         number = as.numeric(number)) %>%
  filter(!field %in% sex) %>%
  filter(!is.na(number))
```

```{r pytabs={'class': 'pytabs-8', 'name': 'R'}}

recipients_year_field_sex %>%
  filter(sex != "All",
         broad_field == "Mathematics and computer sciences") %>%
  ggplot(aes(year, number, color = sex)) +
  geom_line() +
  expand_limits(y = 0) +
  facet_wrap(~ field)

```

```{r pytabs={'class': 'pytabs-9', 'name': 'R'}}

recipients_year_field_sex %>%
  spread(sex, number) %>%
  mutate(pct_male = Male / All) %>%
  filter(broad_field == "Engineering") %>%
  mutate(field = fct_reorder(field, -pct_male)) %>%
  ggplot(aes(year, pct_male, color = field)) +
  geom_line() +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(x = "Year",
       y = "% of PhD recipients reporting as male",
       color = "Major field",
       title = "Breakdown by sex over time within Engineering major fields")

```

```{r pytabs={'class': 'pytabs-10', 'name': 'R'}}

recipients_year_field_sex %>%
  spread(sex, number) %>%
  mutate(pct_male = Male / All) %>%
  filter(broad_field == "Humanities and arts") %>%
  mutate(field = fct_reorder(field, -pct_male)) %>%
  ggplot(aes(year, pct_male, color = field)) +
  geom_line() +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(x = "Year",
       y = "% of PhD recipients reporting as male",
       color = "Major field",
       title = "Breakdown by sex over time within Humanities & Arts major fields")

```

```{r pytabs={'class': 'pytabs-11', 'name': 'R'}}

recipients_year_field_sex %>%
  spread(sex, number) %>%
  mutate(pct_male = Male / All) %>%
  filter(broad_field == "Education") %>%
  mutate(field = fct_reorder(field, -pct_male)) %>%
  ggplot(aes(year, pct_male, color = field)) +
  geom_line() +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(x = "Year",
       y = "% of PhD recipients reporting as male",
       color = "Major field",
       title = "Breakdown by sex over time within Education major fields")
```

Three levels:

* Broad field (Life sciences)
* Major field (Biological and biomedical sciences)
* Subfield (Computational biology)