Online platforms such as GoFundMe have become viable platforms to generate donations for various causes. Research in psychology and behavioral economics have documented a variety of sympathy biases that drive altruistic behavior as function of how a cause is presented. This project seeks to use the latest in machine learning techniques of image analysis and natural language processing to generate quantitative variables from unstructured data such as images and text used to describe the project and link these variables to various relevant outcomes such as amount of funds raised, number of donations, speed of donations.
Requisite Skills and Qualifications:
In the first phase of the project, the RA will help obtain data via web-scraping from an online crowdfunding platform to download the image and text descriptors of projects and outcomes related to projects. The ideal candidate should have experience and be comfortable with programming in Python or R. Any prior experience with web scraping will be a plus.
In the second phase of the project, an interested student can continue working on the project to implement additional machine learning algorithms on the image and text data. However, at this stage priority in selection will be based on interest and necessary skills for Phase ! (webscraping) of the project.