sambasivaraokambala / trump-tweet-analysis

Data and sentiment analysis of Trump's tweets

Analysis of Trump tweets

Data from the Trump Twitter archive.

Versions

A 2017 version of this analysis was used in a Morning Edition radio segment.

The most recent version of this analysis informed an All Things Considered radio story and an NPR.org digital story about President Trump's use of the words "fake" and "phony."

The current iteration has diverged in several significant ways from the original, including:

• Updating the underlying data to cover Trump's presidency to date (through Aug. 31, 2018).
• Updating the analysis code to accommodate data from multiple years.
• Different keyword searches
• Commented out the sentiment analysis

Assumptions

• Python 3 - If you don't have it, you can install it with Homebrew using the command brew install python3
• virtualenvwrapper - Used for isolating notebook dependencies

Bootstrap the project

$ git clone https://github.com/nprapps/trump-tweet-analysis.git
$ mkvirtualenv --python=$(which python3) trump-tweet-analysis
$ pip install -r requirements.txt

Run the notebook

$ cd trump-tweet-analysis
$ workon trump-tweet-analysis
$ jupyter notebook

A listing of files in the project directory should open automatically in your preferred browser. If it doesn't copy the URL given in terminal.

Click on the link for trump-tweets.ipynb to view the notebook.

Click the Cell menu and then the Run All menu item to execute the cells and get the most recent results.