Scraped 5000 Tweets from the #MentalHealthWeek to find the highest frequency of words used.

mentalhealthtweets.Rmd

```{r}
#https://cran.r-project.org/web/packages/rtweet/rtweet.pdf
# install from CRAN
install.packages("rtweet")
install.packages("httpuv")
install.packages("tidyverse")
install.packages("dplyr")
install.packages("tidytext")

# load rtweet
library(rtweet)
library(httpuv)
library(tidyverse)
library(lubridate)
library(ggplot2)
library(dplyr)
library(readr)
library(tidytext)

data("stop_words")
```

```{r}
## setting my permissions, use your own keys here
appname <- "xxxxxxxxxxxx"

## api key (example below is not a real key)
key <- "xxxxxxxxxxxx"

## api secret (example below is not a real key)
secret <- "xxxxxxxxxxxxxx"

## create token named "twitter_token"
twitter_token <- create_token(
  app = appname,
  consumer_key = key,
  consumer_secret = secret)
```

```{r}
tw <- search_tweets("#MentalHealthWeek", n=5000, type = "mixed", include_rts = F)

custom_stop_words <- bind_rows(data_frame(word = c("t.co", "https", "it's", "7", "13", "mentalhealthweek", "amp", "cmha_ntl", "1"), 
                                          lexicon = c("custom")), stop_words)

tidy_mental <- tw %>%
  unnest_tokens(word, text) %>%
  anti_join(custom_stop_words) %>%
  select(word, favorite_count, retweet_count) %>%
  group_by(word) %>%
  count(word, sort = T) %>%
  head(25) %>%
ggplot(aes(x = reorder(word, n), y = n, fill = "blue")) + 
  geom_bar(show.legend = F, stat = "identity", aes(fill = "blue")) + 
  labs(y = "Number of Mentions", x = NULL) +
  ggtitle("Based on 5,000 Tweets from #MentalHealthWeek") +
  coord_flip()

tidy_mental
```