analysis.qmd

---
title: Ethiopian YouTube Channels Statistics
author: Eyayaw
date: today
date-format: long
format:
  html:
    theme:
     light: flatly
     dark : darkly
execute:
  warning: false
  message: false
code-fold: true
code-line-numbers: true
code-link: true
code-tools: true
highlight-style: github
link-external-icon: true
link-external-newwindow: true
---

```{r}
#| label: set up
library(data.table)
library(ggplot2)
library(patchwork)
library(ggtext)
source("R/make_plots.R")
```

```{r}
#| label: import data
channels_info = fread("./data/channels-info.csv")
channels_data = fread("./data/processed/channels-data_merged.csv")
setcolorder(channels_data, c("id", "title"))

# shorten titles
channels_data[, title := gsub("[ -]*\\p{Ethiopic}+[ -]*", "", title, perl = TRUE)]
channels_data[, title := fcase(
  title == "Ethio 360 Media", "Ethio 360",
  title %ilike% "Ethio 251 Media", "Ethio 251",
  title == "Feta Daily News", "Feta Daily",
  title %like% "^EMS", "EMS",
  title %ilike% "Tigra[iy] Media House", "TMH",
  title %ilike% "menelik television 2", "Menelik TV 2",
  rep(TRUE, length(title)), title)
][, title := trimws(title)]

# pick the latest
data_latest = channels_data[retrievedAt == max(retrievedAt), ]
```

```{r}
#| label: plot options
fsize = 12
theme_set(theme_light())
fig_width = 12
fig_height = 0.618 * fig_width
retrieved_at = data_latest$retrievedAt[1]
subtitle = sprintf("Data retrieval time: <span style='color:red;'>%s</span>", retrieved_at)
caption_text = "Author: <span style = 'color:#0090e8;'>twitter.com/eyayaw</span><br>Data: accessed via <span style = 'color:red;'>YouTube Data API v3</span>"
```

```{r}
#| label: chart
#| fig-width: !expr fig_width
#| fig-height: !expr fig_height
#| include: false

lookup = list(
  var = c("viewCount", "subscriberCount", "videoCount", "viewCountPerVideo"),
  label = c("#Views", "#Subscribers", "#Videos", "#Views per video"),
  color = c("#8298f8", "#52b439", "#f68060", "#55bdb4"),
  xupper_lim_expand_mult = c(0.01, .01, .01, .01),
  labeller = c(
    scales::label_number(scale = 10^-6),
    replicate(3, scales::label_comma(scale = 1e-3, suffix = "k"))
  )
)

channels_data_long = data_latest |>
  melt(measure.vars = lookup$var, value.name = "count")

ldata = channels_data_long |> split(by = "variable")
subplots = vector("list", length(ldata))
names(subplots) = names(ldata)

for (.var in names(subplots)) {
  # plot aesthetics
  .aes = with(lookup, list(
   fill = color[match(.var, var)],
   ptitle = label[match(.var, var)],
   xupper_lim_expand_mult = xupper_lim_expand_mult[match(.var, var)],
   labeller = labeller[[match(.var, var)]]
 ))

  # plot
  subplots[[.var]] = ldata[[.var]] |>
    transform(
      title = reorder(title, count), hjust = ifelse(count == max(count), 1, 0)
      ) |>
    ggplot(aes(count, title)) +
    geom_col(fill = .aes$fill, alpha = 0.75, width = 0.9) +
    geom_label(
      aes(label = prettyNum(count, ","), hjust = hjust),
      size = .8 * fsize/.pt, nudge_x = -0.5, fill = "white", label.size = 0) +
    scale_x_continuous(
      #expand = expansion(c(0, .aes$xupper_lim_expand_mult)),
      labels = .aes$labeller
    ) +
    labs(title = .aes$ptitle, x = NULL, y = NULL) +
    coord_cartesian(clip = "off") +
    theme(
      plot.title = element_text(size = fsize, face = "bold", hjust = 0.5),
      axis.text = element_text(size = .9 * fsize),
      panel.grid.major.y = element_blank(),
      panel.grid.minor = element_blank(),
      plot.background = element_rect(fill = "gray99", color = NA)
    )
}


# combine plots ----
panel = wrap_plots(subplots, byrow = TRUE) +
  plot_annotation(
    title = "Stats about Selected Ethiopian YouTube Channels",
    subtitle = subtitle,
    caption = caption_text,
    theme = theme(
      plot.title = element_text(face = "bold", size = 1.5 * fsize),
      plot.subtitle = element_markdown(face = "bold.italic", size = 1.3 * fsize),
      plot.caption = element_markdown(face = "bold.italic", size = .8 * fsize, lineheight = 1.2),
      plot.margin = margin(1, 1, 1, 1, unit = "mm")
    )
  )

print(panel)
```

```{r}
# save figure ----
png_file = sprintf("figs/channels-plot_%s.png", retrieved_at)
ggsave(png_file, panel, ragg::agg_png,
  width = fig_width, height = fig_height, dpi = 300
)
knitr::include_graphics(png_file)
```


```{r}
#| label: channel-views
data_latest |>
  transform(
    title = reorder(title, viewCount),
    hjust = ifelse(viewCount == max(viewCount), 1, 0)
  ) |>
  ggplot(aes(x = viewCount, y = title)) +
  geom_col(fill = "#52b439", alpha = .6) +
  geom_label(
    aes(label = prettyNum(viewCount, big.mark = ","), hjust = hjust),
    size = 10 / .pt, nudge_x = .5, fill = "white", label.size = 0
  ) +
  scale_x_continuous(
    expand = expansion(c(.01, 0.01)),
    labels = scales::label_number(scale = 10^-6)
  ) +
  labs(
    x = "<span>#Views (in 10<sup>6</sup>)</span>", #expression("#Views" ~ ("in" ~ 10^6)),
    y = NULL,
    title = "Total number of views, for a few selected Ethiopian YouTube Channels",
    subtitle = subtitle, caption = caption_text
  ) +
  coord_cartesian(clip = "off") +
  theme_minimal(base_family = "serif") +
  theme(
    legend.position = "none",
    axis.text = element_text(size = .9 * fsize),
    axis.title.x = element_markdown(),
    panel.grid.minor = element_blank(),
    panel.grid.major.y = element_blank(),
    plot.title = element_text(size = fsize, face = "bold"),
    plot.subtitle = element_markdown(size = .9 * fsize, face = "italic"),
    plot.caption = element_markdown(face = "italic", lineheight = 1.2),
    plot.background = element_rect(fill = "gray97", color = NA),
    plot.margin = margin(2, 2, 2, 2, unit = "mm")
  ) -> p_latest

p_latest
```

```{r}
# save figure ----
png_file = sprintf("figs/channels-plot_views_%s.png", retrieved_at)
ggsave(png_file, p_latest, ragg::agg_png, width = 8, height = 6, dpi = 300)
# knitr::include_graphics(png_file)
```

```{r}
scale_exp = -3
axis_label = "subscriber"
data_latest |>
  transform(title = reorder(title, subscriberCount)) |>
  ggplot(aes(subscriberCount, title)) +
  geom_col(fill = "#55bdb4", alpha = .5, width = .8) +
  geom_text(aes(label = scales::label_comma()(subscriberCount)), hjust = -0.05) +
  scale_x_continuous(
    labels = scales::label_number(scale = 10^scale_exp, suffix="k"),
    expand = expansion(c(0.001, 0.12))
    ) +
  labs(
    x = sprintf(
      "<span>#%ss (in 10<sup>%s</sup>)</span>",
      tools::toTitleCase(axis_label), -scale_exp
      ),
    y = NULL,
    title = sprintf(
      "Total number of %ss as of <span style='color:red;'>%s</span>",
      axis_label, retrieved_at
      ),
    caption = sprintf(
"%2$s, and <span>%1$s counts returned by the API are rounded down to<br>three significant figures for %1$s counts greater than 1000 %1$ss.</span>",
axis_label, caption_text)
  ) +
  theme(
    legend.position = "none",
    axis.title.x = element_markdown(),
    plot.title = element_markdown(face = "bold"),
    plot.subtitle = element_markdown(face = "italic"),
    plot.caption = element_markdown(face = "italic", hjust = 0, lineheight = 1.2, margin = margin(r=4)),
    plot.margin = margin(t=2, r=2, b=2, l=2),
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank()
  )

# save figure ----
png_file = sprintf("figs/channels-plot_subscribers_%s.png", retrieved_at)
ggsave(png_file, last_plot(), ragg::agg_png, width = 8, height = 6, dpi = 300)
#knitr::include_graphics(png_file)

```



```{r}
# bar_plot(data_latest, viewCount, title, scales::label_comma()(viewCount))
```


```{r}
weekly_ts_plot <-
  function(data, date_var, series_var, ..., geom.point.pars = list(), geom.line.pars = list()) {
    default.line.pars = list(linewidth = 1.01, color = "#55bdb4")
    default.point.pars = list(size = 9 / .pt, color = "#55bdb4")
    ggplot(data, mapping = aes(x = {{ date_var }}, y = {{ series_var }}, ...)) +
      do.call("geom_line", modifyList(default.line.pars, geom.line.pars)) +
      do.call("geom_point", modifyList(default.point.pars, geom.point.pars)) +
      xlab("Date") +
      scale_y_continuous(labels = scales::label_number(scale = 1e-3, suffix = "k")) +
      scale_x_date(
        date_breaks = "2 days",
        date_labels = "%b %d",
        expand = expansion(c(0.01, 0.01))
      )
  }
```

```{r}
 p = channels_data |>
  subset(retrievedAt >= "2023-01-01") |>
  _[order(retrievedAt, title),
    .(retrievedAt = as.Date(retrievedAt), daily_views = viewCount - shift(viewCount)), title
   ] |>
  weekly_ts_plot(retrievedAt, daily_views) +
  ylab(expression(Delta ~ "Total Views"))
```



```{r}
#| label: table
data_latest[, .(title, viewCount, subscriberCount, videoCount, videoCount, publishedAt, age)] |>
  kableExtra::kable(booktabs = TRUE, fontsize = 10) |>
  kableExtra::kable_styling(bootstrap_options = c("striped", "bordered"))
```