Using emoji with Highcharter

Emoji are text, categories and tiny pictures at the same time. That makes them surprisingly useful marks for an interactive visualization.

This historical experiment originally used a live sample of posts containing #rstats. The aggregated observations have been recovered from the published Highcharter widgets, so the article now renders without an API or account.

library(dplyr)
library(highcharter)
library(readr)

emoji_frequency <- read_csv(
  "data/emoji-frequency.csv",
  show_col_types = FALSE
)

emoji_engagement <- read_csv(
  "data/emoji-engagement.csv",
  show_col_types = FALSE
)

Emoji as categories

The first view simply counts each symbol. Highcharts accepts the emoji as an ordinary category label; no custom icon files are needed.

hchart(
  emoji_frequency,
  "bar",
  hcaes(emoji, n),
  name = "Count"
) |>
  hc_title(text = "Most frequent emoji in the historical #rstats sample") |>
  hc_xAxis(
    min = 0,
    max = 30,
    scrollbar = list(enabled = TRUE)
  ) |>
  hc_yAxis(title = list(text = "Count", align = "high")) |>
  hc_tooltip(
    headerFormat = "{point.key}",
    pointFormat = " {point.y}"
  ) |>
  hc_credits(enabled = FALSE) |>
  hc_size(height = 700)

Emoji as markers

The second view compares accumulated favorites and retweets. The scatter points are invisible circles; data labels place the corresponding emoji at each coordinate instead.

hchart(
  emoji_engagement,
  "scatter",
  hcaes(favorite_count, retweet_count),
  name = "Emoji",
  marker = list(radius = 0),
  dataLabels = list(
    enabled = TRUE,
    format = "{point.emoji}",
    allowOverlap = TRUE,
    style = list(fontSize = "20px", textOutline = "none"),
    y = 20
  )
) |>
  hc_title(text = "Emoji engagement in the historical #rstats sample") |>
  hc_xAxis(
    type = "logarithmic",
    title = list(text = "Favorites"),
    minRange = 1
  ) |>
  hc_yAxis(
    type = "logarithmic",
    title = list(text = "Retweets"),
    minRange = 1
  ) |>
  hc_chart(zoomType = "xy") |>
  hc_tooltip(
    useHTML = TRUE,
    backgroundColor = "rgb(244, 246, 248)",
    borderWidth = 0,
    pointFormat = paste0(
      "<b style='font-size:22px'>{point.emoji}</b><br>",
      "Favorites: {point.x}<br>Retweets: {point.y}"
    )
  ) |>
  hc_credits(enabled = FALSE)

The exact rankings belong to one historical sample, but the graphical idea is still useful: when a symbol is already meaningful to the reader, it can act as both label and mark.

Recovery note

Both local CSV files were recovered from the JSON embedded in the original widgets. The render performs no social-media queries.