Scrape only Tweets with Emojis in R – Twitter API

Note: I assume you’re not looking for all emoji, since they include quite common characters:

(from https://unicode.org/Public/UNIDATA/emoji/emoji-data.txt)

Unicode library

To get the Unicode block for one or more characters, we can use the Unicode library:

library("Unicode") # install.packages("Unicode")

A few examples:

> "😎" |> utf8ToInt() |> u_char_properties("Block")

            Block
U+1F60E Emoticons

> "👨‍👨‍👦" |> utf8ToInt() |> u_char_properties("Block")

                                  Block
1 Miscellaneous Symbols and Pictographs
2                   General Punctuation
3 Miscellaneous Symbols and Pictographs
4                   General Punctuation
5 Miscellaneous Symbols and Pictographs

> "🤐" |> utf8ToInt() |> u_char_properties("Block")

                                       Block
U+1F910 Supplemental Symbols and Pictographs

> "✅" |> utf8ToInt() |> u_char_properties("Block")

          Block
U+2705 Dingbats

> "☝️" |> utf8ToInt() |> u_char_properties("Block")

                       Block
U+261D Miscellaneous Symbols
U+FE0F   Variation Selectors

> "☎️" |> utf8ToInt() |> u_char_properties("Block")

                       Block
U+260E Miscellaneous Symbols
U+FE0F   Variation Selectors

> "♍" |> utf8ToInt() |> u_char_properties("Block")

                       Block
U+264D Miscellaneous Symbols

> "🫃🏽" |> utf8ToInt() |> u_char_properties("Block")

                                        Block
U+1FAC3    Symbols and Pictographs Extended-A
U+1F3FD Miscellaneous Symbols and Pictographs

> "🚂" |> utf8ToInt() |> u_char_properties("Block")

                            Block
U+1F682 Transport and Map Symbols

Matching all emoji-like characters could be done like this:

blocks <- c("Emoticons",
            "Miscellaneous Symbols and Pictographs",
            "Supplemental Symbols and Pictographs",
            "Dingbats",
            "Miscellaneous Symbols",
            "Symbols and Pictographs Extended-A",
            "Transport and Map Symbols")

> "😎" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] TRUE

> "👨‍👨‍👦" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] TRUE

> "🤐" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] TRUE

> "☎️" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] TRUE

> "♍" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] TRUE

> "🫃🏽" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] TRUE

> "🚂" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] TRUE

> "#" |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0
[1] FALSE

Integrating into your code

library(rtweet)
library(dplyr)

get_token() # Connects with Twitter API
Uber <- search_tweets("uber", n = 100, lang = "en")

Uber_filtered <- Uber %>%
  rowwise() %>%
  filter(text |> utf8ToInt() |> u_char_properties("Block") |> unlist() |> intersect(blocks) |> length() > 0)