We plan to start by harvesting tweets from Twitter accounts of two categories: UT Administration/Faculty/Staff and UT students/general population. The textual sentiment of these tweets will be calculated and compared during pre and post COVID-19 time periods, and we will compare the general attitudes of the tweets from these group's social media presence. Further comparison could include searching for correlation between sentiment and COVID-19 rates and other metrics. Also, comparison of this data between universities and looking at word frequencies could potentially lead to interesting findings. We will initially use a scraper to avoid Twitter's weak API for data gathering, and plan on using Jupyter Notebooks with numerous of the powerful data science libraries available for use in Python (NumPy, Pandas, Matplotlib, SciPy, TensorFlow, etc).
- Tucker Miles
- Vicki Tang