Francesca T
- Research Program Mentor
PhD at Princeton University
Expertise
Applied statistics, data science, political science, social science, machine learning, regression
Bio
I am a statistician by training but working on political science problems using statistical and machine learning methods, as I have always been passionate about the intersection of statistics and social science. I am currently a postdoctoral research associate at the Department of Politics at Princeton University and I received my Ph.D. in Operations Research and Financial Engineering from Princeton University in 2022, where I was advised by Professor Jianqing Fan. Before that, I received my bachelor's in mathematics from Duke University in 2017. Outside of research, I co-founded a startup Leap Careers LLC, am a huge soccer fan (Manchester City!), am extremely enthusiastic about languages (I speak 3), read a ton of books, and spend a lot of time cooking and baking sourdough bread.Project ideas
How does Twitter impact stock prices?
In this project, you will use the Twitter API to scrape Tweets to see if there's a direct impact on certain movements in stock prices. For example, if there was a lot of activity surrounding a certain company or hashtag, perhaps we see increases or decreases in certain stock prices immediately after. You may be able to find very interesting links in how Twitter activity can move the market and learn about how to apply simple correlation and linear models. This project could be written as a scientific research paper.
Text Analysis in Political Science
A review paper or essay on how has the use of text analysis and large language models progressed in the field of political science in the last decade or so. With the advancement of NLPs and machine learning, you can explore how much political scientists have taken advantage of language models to solve various problems. In the process, you will learn about different NLP fields such as sentiment analysis, topic modeling, word embedding, text summarization, etc. In addition, how these different techniques have been used to answer political science questions.
Demographics and Voting
Using linear models, you will investigate how demographic variables such as population, age, etc. are connected to voting behavior. For example, do counties with larger population actually vote more for the Democratic Party? Do states with more college educated people or diverse populations vote less for the Republican Party? You will explore all this with linear regression models and either write a research paper or put together a large poster of your findings.