2021-01-12 artists.Rmd

---
title: "R Notebook"
output: html_notebook
---

```{r}
library(tidyverse)
library(tidyr)
library(ggplot2)
library(plotly)
library(ggrepel)
library(ggdark)
library(ggtext)
library(patchwork)

artists <- data.table::fread("~/github/tidytues/tidytuesday/data/2021/2021-01-12/artists.csv") %>% 
  mutate(gender = ifelse(is.na(gender), "Unknown", gender),
         gender = factor(gender,levels = c("Male","Female", "Unknown"), ordered = TRUE),
         life_yr = yearOfDeath - yearOfBirth,
         pre_1850 = yearOfBirth < 1850,
         name_len = stringr::str_length(name) - 2) %>% 
  separate(col = "placeOfBirth", sep = ",", into = c("birth_city", "birth_country"), remove = F) %>% 
  separate(col = "placeOfDeath", sep = ",", into = c("death_city", "death_country"), remove = F) %>% 
  mutate(moved_countries = birth_country != death_country,
         birth_country = ifelse(is.na(birth_country), "Unknown", birth_country),
         death_country = ifelse(is.na(death_country), "Unknown", death_country))
artwork <- data.table::fread("~/github/tidytues/tidytuesday/data/2021/2021-01-12/artwork.csv") %>% 
  mutate(artistRole = as.factor(artistRole),
         medium = as.factor(medium),
         units = as.factor(units),
         area = width * height,
         title_len = stringr::str_length(title)) %>% 
  left_join(y = artists %>% select(name, gender, yearOfBirth, birth_city, birth_country, life_yr), by = c("artist"="name"))
artists
artwork
```

```{r fig.height=5, fig.width=7}
artists %>% 
  ggplot(aes(x = life_yr, fill = pre_1850)) +
  geom_histogram() +
  facet_grid(~gender) + 
  ggdark::dark_theme_classic()

artists %>% 
  ggplot(aes(y = life_yr, x = yearOfBirth, color = gender)) +
  geom_point(alpha = 0.75) +
  geom_smooth() + 
  ggdark::dark_theme_classic()

artists %>% 
  ggplot(aes(y = moved_countries, x = yearOfBirth, color = gender)) +
  geom_point(alpha = 0.75) +
  ggdark::dark_theme_classic()

```

```{r fig.height=4, fig.width=15}
artists %>% 
  select(year = yearOfBirth, moved_countries, yearOfBirth) %>% 
  group_by(year) %>% 
  count(moved_countries) %>% 
  # tidyr::pivot_wider(names_from = moved_countries, values_from = n) %>% 
  ggplot(aes(x = year, y = n, color = moved_countries)) +
  geom_line() +
  ggdark::dark_theme_classic()
```

```{r fig.height=4, fig.width=15}
ggplot(artists) +
  # geom_point(aes(x = yearOfBirth, y = life_yr), color = "gray30", alpha = 0.5, size = 2) +
  # geom_point(aes(x = yearOfDeath, y = life_yr), color = "gray30", alpha = 0.5, size = 2) +
  geom_segment(aes(x = yearOfBirth, xend = yearOfDeath, y = life_yr, yend = life_yr, color = gender), alpha = 0.7) +
  ggdark::dark_theme_classic()
```

```{r fig.height=5, fig.width=6}
artists %>% 
  group_by(birth_country, gender) %>% 
  count() %>%
  group_by(birth_country) %>% 
  mutate(n_sum = sum(n)) %>% 
  ungroup() %>%
  top_n(20, n_sum) %>% 
  ggplot() +
  geom_col(aes(y = forcats::fct_reorder(birth_country, n_sum), x = n, fill = gender)) +
  # scale_x_log10() + 
  ggdark::dark_theme_classic()
```
```{r fig.height=12, fig.width=10}
artists %>% 
  group_by(birth_country, gender) %>% 
  count() %>% 
  tidyr::pivot_wider(names_from = "gender", values_from = "n") %>%
  tidyr::replace_na(replace = list(Female = 0, Male = 0, Unknown = 0)) %>% 
  mutate(pc_fem = Female / (Male + Female + Unknown),
         total = Female + Male + Unknown) %>% # View() 
  tidyr::drop_na(pc_fem) %>%
  ggplot() + 
  geom_point(aes(x = forcats::fct_reorder(birth_country, pc_fem), y = pc_fem, size = total), color = "gray70", alpha = 0.8) +
  # geom_point(aes(x = forcats::fct_reorder(birth_country, pc_fem), y = pc_fem, size = Female), color = "forestgreen", alpha = 0.8) +
  scale_size(range = c(1, 12)) + 
  # scale_size(limits = c(0, 1000), range = c(0, 10)) + 
  # geom_boxplot() + 
  ggdark::dark_theme_classic() +
  theme(#axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5),
        axis.line = element_blank(),
        panel.grid.major.x = element_line(colour = "gray10")) + 
  scale_y_continuous(labels = scales::label_percent()) + 
  coord_flip()
  # theme(legend.position = "none")
```
```{r fig.height=5, fig.width=12}
artists %>% 
  group_by(yearOfBirth, birth_country, gender) %>% 
  count() %>% 
  tidyr::pivot_wider(names_from = "gender", values_from = "n") %>% 
  mutate(pc_fem = Female / (Male + Female)) %>% 
  tidyr::drop_na(pc_fem) %>%
  ggplot(aes(x = yearOfBirth, y = pc_fem, color = birth_country)) + 
  geom_point() + 
  ggdark::dark_theme_classic() #+ 
  # theme(legend.position = "none")
```

```{r fig.height=7, fig.width=9}
artists %>% 
  separate(name, sep = "[,\\s].", into = c("lname", "fname", "mname"), extra = "merge") %>%   
  mutate(fname_len = stringr::str_length(fname)) %>% 
  group_by(gender, yearOfBirth) %>% 
  summarise(med_len = median(fname_len, na.rm = TRUE)) %>% 
  ggplot(aes(x = yearOfBirth, y = med_len, color = gender)) + 
  # geom_point(alpha = 0.4, size = 2) +
  geom_smooth(se = F, show.legend = F) + 
  ggdark::dark_theme_classic() +
  scale_color_manual(values = list("Male" = "#e5e5e5", "Female" = "#f4a261", "Unknown" = "#e76f51"))
```

```{r}
artwork %>% 
  skimr::skim()
```
```{r}
# devtools::install_github('Ather-Energy/ggTimeSeries')
library(ggTimeSeries)
```

```{r}
medium_dat <-
  artwork %>%
  count(year, medium) %>% 
  pivot_wider(names_from = medium, values_from = n) %>% 
  pivot_longer(-year) %>% 
  tidyr::replace_na(list(value = 0))
medium_dat
```
```{r}
medium_dat %>% 
  group_by(name) %>% 
  summarise(v = sum(value)) %>% 
  top_n(100, v) %>% 
  arrange(-v)
```

```{r}
medium_cleaned <- medium_dat %>% 
  mutate(medium_cleaned = case_when(
    grepl(pattern = "Graphite", x = name) ~ "Graphite",
    grepl(pattern = "Oil paint", x = name) ~ "Oil Paint",
    grepl(pattern = "Screenprint", x = name) ~ "Screenprint",
    grepl(pattern = "Watercolour", x = name) ~ "Watercolour",
    grepl(pattern = "Ink on paper", x = name) ~ "Ink",
    grepl(pattern = "photograph|Photograph", x = name) ~ "Photograph",
    grepl(pattern = "engraving|Engraving", x = name) ~ "Engraving",
    grepl(pattern = "chalk|Chalk", x = name) ~ "Chalk",
    name == "NA" ~ "Other"
  )) %>% 
  filter(!is.na(medium_cleaned))
```

```{r}
medium_cleaned %>% 
  ggplot(aes(x = year, y = value, fill = medium_cleaned)) + 
  geom_area()
```
```{r fig.height=8, fig.width=12}
make_plot <- function(medium_dat,
                      yr_max = 1900,
                      num_years = 100) {
  yr_min <- yr_max - num_years
  
  dat <- medium_dat %>%
    filter(year > yr_min, year < yr_max)
  
  # top_mediums <-  dat %>%
  #   group_by(name) %>%
  #   summarise(n = sum(value)) %>%
  #   filter(name != "NA") %>%
  #   top_n(n = 5, wt = n) %>%
  #   arrange(-n) %>% 
  #   pull(name)
  
  dat %>%
    # filter(name %in% top_mediums) %>%
    mutate(medium_cleaned = factor(medium_cleaned, levels = medium_cleaned)) %>% 
    ggplot(aes(x = year, y = value)) +
    stat_steamgraph(aes(group = medium_cleaned, fill = name)) +
    # theme_linedraw() +
    scale_x_continuous(breaks = seq(yr_min, yr_max, 10)) + 
    theme(legend.title = element_blank(),
          axis.line = element_line(size = 0),
          axis.line.x.top = element_line(size = 0),
          plot.background = element_rect(fill = "white"),
          panel.background = element_rect(fill = "white"),
          axis.text.y = element_blank(),
          axis.ticks.y = element_blank(),
          panel.grid.major.x = element_line(linetype = "dotted", size = 0.3, colour = "gray60")
          ) +
    labs(x = "", y = "")
}

p1 <- make_plot(medium_cleaned, 1860, 100)
p2 <- make_plot(medium_cleaned, 2010, 100)

library(cowplot)
cowplot::plot_grid(p1, p2, nrow = 2, align = "hv")

```

```{r fig.height=8, fig.width=12}
make_plot <- function(medium_dat,
                      yr_max = 1900,
                      num_years = 100) {
  yr_min <- yr_max - num_years
  
  dat <- medium_dat %>%
    filter(year > yr_min, year < yr_max)
  
  top_mediums <-  dat %>%
    group_by(name) %>%
    summarise(n = sum(value)) %>%
    filter(name != "NA") %>%
    top_n(n = 5, wt = n) %>%
    arrange(-n) %>% 
    pull(name)
  
  dat %>%
    filter(name %in% top_mediums) %>%
    mutate(name = factor(name, levels = top_mediums, labels = stringr::str_trunc(top_mediums, width = 25))) %>% 
    ggplot(aes(x = year, y = value)) +
    stat_steamgraph(aes(group = name, fill = name)) +
    # theme_linedraw() +
    scale_x_continuous(breaks = seq(yr_min, yr_max, 10)) + 
    theme(legend.title = element_blank(),
          axis.line = element_line(size = 0),
          axis.line.x.top = element_line(size = 0),
          plot.background = element_rect(fill = "white"),
          panel.background = element_rect(fill = "white"),
          axis.text.y = element_blank(),
          axis.ticks.y = element_blank(),
          panel.grid.major.x = element_line(linetype = "dotted", size = 0.3, colour = "gray60")
          ) +
    labs(x = "", y = "")
}

p1 <- make_plot(medium_dat, 1860, 100)
p2 <- make_plot(medium_dat, 2010, 100)

library(cowplot)
cowplot::plot_grid(p1, p2, nrow = 2, align = "hv")

```
```{r fig.height=5, fig.width=12}
make_plot <- function(medium_dat,
                      yr_max = 1900,
                      num_years = 100) {
  yr_min <- yr_max - num_years
  
  dat <- medium_dat %>%
    filter(year > yr_min, year < yr_max)
  
  top_mediums <-  dat %>%
    group_by(name) %>%
    summarise(n = sum(value)) %>%
    filter(name != "NA") %>%
    top_n(n = 5, wt = n) %>%
    arrange(-n) %>% 
    pull(name)
  
  dat %>%
    filter(name %in% top_mediums) %>%
    mutate(name = factor(name, levels = top_mediums)) %>% 
    ggplot(aes(x = year, y = value)) +
    geom_area(aes(fill = name)) +
    theme_bw() +
    scale_x_continuous(breaks = seq(yr_min, yr_max, 10))
}

p1 <- make_plot(medium_dat, 1860, 100)
p2 <- make_plot(medium_dat, 2010, 100)

p1 / p2
```

```{r}
gender_dat <-
  artwork %>% select(year, gender) %>% 
  filter(gender != "Unknown") %>% 
  tidyr::drop_na() %>% 
  group_by(year) %>% 
  count(gender) %>% 
  pivot_wider(names_from = gender, values_from = n) %>% 
  pivot_longer(-year) %>% 
  tidyr::replace_na(list(value = 0))
gender_dat

dat <- gender_dat %>% 
  filter(year > 1750, year < 1850)

dat %>% 
  ggplot(aes(x = year, y = value)) + 
  stat_steamgraph(aes(group = name, fill = name)) +
  theme_bw() + 
  scale_x_continuous(breaks = seq(1750, 1850, 10))-> p1

dat <- gender_dat %>% 
  filter(year >= 1950, year <= 2010)

dat %>% 
  ggplot(aes(x = year, y = value)) + 
  stat_steamgraph(aes(group = name, fill = name)) +
  theme_bw() + 
  scale_x_continuous(breaks = seq(1950, 2010, 10))-> p2

library(patchwork)
p1 / p2
```

```{r}
title_len_dat <-
  artwork %>% select(year, gender, area) %>% 
  filter(gender != "Unknown") %>% 
  tidyr::drop_na() %>% 
  group_by(year, gender) %>% 
  summarise(mean_title_len = median(area)) %>% 
  pivot_wider(names_from = gender, values_from = mean_title_len) %>% 
  pivot_longer(-year) %>% 
  tidyr::replace_na(list(value = 0))
title_len_dat

dat <- title_len_dat %>% 
  filter(year > 1750, year < 1850)

dat %>% 
  ggplot(aes(x = year, y = value)) + 
  stat_steamgraph(aes(group = name, fill = name)) +
  theme_bw() + 
  scale_x_continuous(breaks = seq(1750, 1850, 10))-> p1

dat <- title_len_dat %>% 
  filter(year >= 1900, year <= 2010)

dat %>% 
  ggplot(aes(x = year, y = value)) + 
  stat_steamgraph(aes(group = name, fill = name)) +
  theme_bw() + 
  scale_x_continuous(breaks = seq(1900, 2010, 10))-> p2

library(patchwork)
p1 / p2
```