TweetScreeneR

A Shiny Application to manually screen — or classify — Tweets downloaded from the Twitter API with the rtweet-package.

The app is currently being actively developed towards the first beta release. There are no known issues or bugs, but the usability is limited: only dichotomous classification (Include/Exclude) is possible, with no option to customize the labels.

Usage

1. Place your .RDS files containing the tweets in a folder and point TweetScreeneR to that folder. All files starting with the string 'tweets' will be imported.
2. Use the input controls to filter based on date
3. Subset the tweets by searching (with regular expressions)
4. Anonymize @mentions if necessary
5. Start screening!

Every time the 'Start'-Button is pressed a session file is created in the output folder (default: data. This can be changed under advanced settings in the UI). The output is a .csv-file containing the tweet text, ID and the decisions. By checking the box 'Remove already screened tweets' only tweets that haven't been screened/classified are shown in subsequent sessions.

TODO

• Add user feedback for chosen decisions
• Custom labels based on user input
  • Option to save the configuration in a file
• Option to change the file format and naming convention of imported and exported files (currently only the folder can be changed)
  • Option to choose exported fields