Why Do Songs, Books, and Movies Become Popular?

Mentor Areas

Description:

Why do some songs, books and movies catch on and become popular while others fail? Why do some online articles suck us in and get lots of engagement while others don’t? We’re interested in using natural language processing and textual analysis to help answer these questions.  Ongoing projects involve analyzing song lyrics to predict Billboard rankings, analyzing movies scripts to plot the emotional arc of narratives and predict ratings and ticket sales, and analyzing online content to understand why certain articles get longer versus shorter reads. Student will help analyze data and eventually generate new theories to test.  Open to any year but preferences will be given to upper level students with more experience.  We would like to have a student who can begin in the Spring semester of 2018!

Preferred Qualifications

While no prior experience is required, and anyone interested should email, ideal candidates will have:

• Experience programming in Python and R, especially with processing large amounts of text data. Experience in one or more of the following packages: Pandas, seaborn, NLTK, spaCy, numpy, scipy, scikit-learn, and statsmodels or their R counterparts (dplyr, ggplot, tidytext, etc.)
• Coursework in one or more of the following, or similar courses: statistics (STAT 417, 476), machine learning (CIS 519, 520, 521), computational linguistics (CIS 530), linguistics (LING 449)
• Bonus if you have:
  • Experience with jupyter notebooks, for prototyping, exploratory data analysis, and reporting
  • Experience in sentiment analysis and/or automated assessment of text readability/quality.
  • Bash scripting (e.g., for computing on Wharton’s High Performance Computing Cluster)
  • Git for version control