For Dickinson College Student/Faculty Collaborative Research, I was able to scrape Google Reviews data from real restaurants to create insights on the restaurants market, alongside with discovering relationships between certain reviewer statistics that may help predict ratings from future reviewers. By utilizing OpenAI's gpt-4-turbo model, I compressed the large body of review texts into smaller "representative" chunks using a machine learning model. This model allows people to receive a summarized text review of any restaurant they are interested in to obtain a better understanding of what the restaurant has to offer, and if it was perceived well by previous customers.
@ author: Michael Freda @ Advised by Professor Eren Bilen