So I’ve just about beaten this Twitter mining thing to death but I’ve made a couple improvements that I think make another update worthwhile:

1. I’ve built out the R Shiny App a little more to include some cool time-series analyzers for hashtag engagement
2. I’ve created an R-Markdown file that details the end-to-end flow of the project
3. I’ve organized the whole thing in a GitHub repository for your collective enjoyment

Project Recap

I’ve established a link to Twitter’s API using the rtweet package which I use to grab tweets from and about NOAA Fisheries. I’ve also created database tables in a SQL Server database to store these tweets. I refresh the database once-a-week by pulling the most recent 7 days worth of tweets and pushing them to the database. From there a run a script that updates the data called on by the Shiny App.

The Action

The GitHub Repository with all the code for the end-to-end project is here.

The interactive R Shiny App is here.

R Skills

1. Use RODBC to set up and manage the SQL Server database using R.
2. Use rtweet to search for and download tweets according to a search criteria
3. Use R Shiny to display tweets interactively in an app
4. Use rsconnect to web deploy the app from the shinyapp.io server.