Safeguarding Voters from State Sponsored Content
We were inspired by the challenge offered by Leidos, "Best Use of Machine Learning to Detect Sponsored Content", and wanted to offer a unique approach to tackling sponsored content. Thus, we decided to attempt to detect state-sponsored content specifically. We utilized many features of Python's natural language tool kit (NLTK) to comprehensively analyze over 200,000 known Russian sponsored tweets in order to gain insight into how state-sponsored, specifically Russian sponsored, twitter content differs from regular, non-sponsored political tweets and built various Naive Bayes Machine Learning Algorithms to attempt to classify tweets as either sponsored or non-sponsored. We illustrated our findings using Plotly.py, a package for creating and sharing interactive, web-based data visualizations.