rdemo1.R

# Demo 1: Rtweet en word clouds
# install.pacakges("rtweet")

library(tidyverse)
library(rtweet)
library(quanteda)
library(quanteda.textplots)
library(RColorBrewer)

auth_setup_default()

tweets = search_tweets("ukraine", n = 1000, include_rts = FALSE)

tweets |> 
  corpus() |>
  tokens() |> 
  dfm() |>
  textplot_wordcloud(max_words=100)

tweets |> 
  filter(lang == "en") |>
  corpus() |>
  tokens() |> 
  dfm()  |> 
  dfm_remove(min_nchar=2) |>
  dfm_remove("ukraine") |>
  dfm_remove(stopwords()) |>
  textplot_wordcloud(max_words=200, color=brewer.pal(8, "Dark2"),
                     random_color=TRUE, random_order = TRUE)


# Demo 2: Tidyverse
d = read_csv("https://raw.githubusercontent.com/houstondatavis/data-jam-august-2016/master/csv/county_facts.csv")

# Filter, select, mutate
states = d |> filter(is.na(state_abbreviation), fips != 0) |>
  select(fips, area_name, population=Pop_2014_count, pop_change=Pop_change_pct, 
         white=Race_white_pct, college=Pop_college_grad_pct, income=Income_per_capita) |>
  mutate(growing=pop_change>1) 

ggplot(states) + geom_point(aes(x=college, y=income))
ggplot(states) + geom_point(aes(x=college, y=income, color=growing))
ggplot(states) + geom_point(aes(x=college, y=income, size=population, color=white), alpha=.5)