african-american-history.Rmd

---
title: "African-American History"
output: html_document
---

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

```{r}
library(tidyverse)
library(tidytuesdayR)
library(scales)
theme_set(theme_light())

tuesdata <- tt_load('2020-06-16')
```

```{r}
slave_routes <- tuesdata$slave_routes

slave_routes %>%
  summarize(sum(n_slaves_arrived, na.rm = TRUE))

slave_routes %>%
  count(ship_name, sort = TRUE)

slave_routes %>%
  ggplot(aes(n_slaves_arrived)) +
  geom_histogram()
```

```{r}
slave_routes %>%
  filter(!is.na(ship_name), !is.na(n_slaves_arrived)) %>%
  filter(fct_lump(ship_name, 12) != "Other") %>%
  mutate(ship_name = fct_reorder(ship_name, n_slaves_arrived)) %>%
  ggplot(aes(n_slaves_arrived, ship_name)) +
  geom_boxplot()
```

```{r}
slave_routes %>%
  count(port_arrival, sort = TRUE)

slave_routes %>%
  count(port_origin, sort = TRUE)


slave_routes %>%
  ggplot(aes(year_arrival)) +
  geom_histogram() +
  geom_vline(xintercept = c(1776, 1865), lty = 2) +
  labs(title = "# of slave ships over time")
```

```{r}
slave_routes %>%
  filter(!is.na(port_origin)) %>%
  mutate(port_origin = str_trunc(port_origin, 25),
         port_origin = fct_lump(port_origin, 24),
         port_origin = fct_reorder(port_origin, year_arrival)) %>%
  count(port_origin,
        decade = 10 * (year_arrival %/% 10)) %>%
  ggplot(aes(decade, n)) +
  geom_line() +
  facet_wrap(~ port_origin) +
  geom_vline(xintercept = c(1807), lty = 2) +
  theme(strip.text = element_text(size = 6))

slave_routes %>%
  filter(port_origin == "New York") %>%
  arrange(desc(year_arrival)) %>%
  View()
```

```{r}
slave_routes %>%
  filter(fct_lump(ship_name, 12) != "Other") %>%
  count(ship_name, port_origin, port_arrival, sort = TRUE) %>%
  group_by(ship_name) %>%
  mutate(percent = n / sum(n)) %>%
  top_n(1, percent) %>%
  arrange(desc(percent))
```

```{r}
library(grid)
library(ggraph)

set.seed(2020)

slave_routes %>%
  count(port_origin, port_arrival, sort = TRUE) %>%
  filter(port_origin != port_arrival) %>%
  head(40) %>%
  ggraph(layout = "fr") +
  geom_edge_link(aes(width = n),
                 arrow = arrow(type = "closed",
                               length = unit(.1, "inch"))) +
  geom_node_point() +
  geom_node_text(aes(label = name), repel = TRUE) +
  scale_edge_width_continuous(range = c(.5, 4)) +
  labs(title = "Common slave routes",
       edge_width = "# of ships")
```

```{r}
slave_routes %>%
  group_by(port_origin) %>%
  summarize(n_ships = n(),
            total_recorded = sum(n_slaves_arrived, na.rm = TRUE),
            pct_missing = mean(is.na(n_slaves_arrived)),
            estimated_total = mean(n_slaves_arrived, na.rm = TRUE) * n()) %>%
  arrange(desc(n_ships))
```

```{r}
census <- tuesdata$census

census_gathered <- census %>%
  mutate(other = total - white - black) %>%
  arrange(year) %>%
  gather(racial_category, population, white, black_free, black_slaves, other)

census_gathered %>%
  filter(region == "USA Total") %>%
  ggplot(aes(year, population, fill = racial_category)) +
  geom_col() +
  scale_y_continuous(labels = comma) +
  labs(x = "Year",
       y = "Census Population",
       fill = "Racial category",
       title = "Census racial makeup of US, 1790-1870",
       subtitle = "No 'other' category existed before 1860")
```

```{r}
census_gathered %>%
  filter(region != "USA Total", !is.na(division)) %>%
  mutate(division = fct_reorder(division, -population, sum)) %>%
  ggplot(aes(year, population, fill = racial_category)) +
  geom_col() +
  scale_y_continuous(labels = comma) +
  facet_wrap(~ division) +
  labs(x = "Year",
       y = "Census Population",
       fill = "Racial category",
       title = "Census racial makeup of US, 1790-1870",
       subtitle = "No 'other' category existed before 1860")
```

```{r}
census_gathered %>%
  filter(region != "USA Total", !is.na(division)) %>%
  mutate(division = fct_reorder(division, -population, sum)) %>%
  group_by(division, year) %>%
  mutate(percent = population / sum(population)) %>%
  ggplot(aes(year, percent, fill = racial_category)) +
  geom_col() +
  scale_y_continuous(labels = percent) +
  facet_wrap(~ division) +
  labs(x = "Year",
       y = "% of Census Population",
       fill = "Racial category",
       title = "Census racial makeup of US, 1790-1870",
       subtitle = "No 'other' category existed before 1860")
```

```{r}
library(ggwordcloud)

name_counts <- tuesdata$african_names %>%
  count(name, gender, sort = TRUE)

wordcloud(name_counts$name, name_counts$n)

name_counts %>%
  head(100) %>%
  ggplot(aes(label = name, size = n, color = gender)) +
  geom_text_wordcloud()
```

```{r}
library(tidytext)

tuesdata$african_names %>%
  filter(!is.na(gender)) %>%
  mutate(gender = fct_recode(gender, Man = "Boy", Woman = "Girl")) %>%
  count(name, gender, sort = TRUE) %>%
  group_by(gender) %>%
  top_n(20, n) %>%
  ungroup() %>%
  mutate(name = reorder_within(name, n, gender)) %>%
  ggplot(aes(n, name)) +
  geom_col() +
  scale_y_reordered() +
  facet_wrap(~ gender, scales = "free")
```

```{r}
tuesdata$african_names %>%
  count(country_origin, sort = TRUE)
```