Our lab, iDPI, researches social media and builds new tools designed to address some of the persistent problems with online interaction. We are especially interested in the ways people in different cultures use the same platforms in very different ways. Much of our current research focuses on comparing how people in different nations use shortform video on TikTok and YouTube.
Project Description
We're building a set of tools that can help us determine the format videos are in, and whether they've been uploaded from other platforms - for instance, is a video on YouTube previously posted on TikTok. We'd love to work with people with machine learning and image recognition skills. We also have several projects focused on language transcription and hand-coding of videos in different languages - we're excited to work with multilingual students.
Learning Objectives:
Students will understand how a research analysis pipeline is developed and used to characterize large collections of videos.
Students will be able to create new tools for use within that pipeline and conduct analysis of large video corpora.
Skills to learn:
Ability to develop software in a Unix environment - Mac is fine, but folks working only on PC have had a hard time on this project previously.
Some ML knowledge, especially around image recognition. Knowledge of video - especially ffmpeg - is a big plus.