• Stars
    star
    1
  • Language
    Python
  • License
    MIT License
  • Created almost 3 years ago
  • Updated almost 3 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

How can brands analyze Internet memes to identify critiques or compliments about products or services? And how should they respond to those memes? These are the two research questions that I have chosen for my bachelor's degree thesis. To perform this research, I have used the PushShift API to develop a dataset of 17,000 brand memes retrieved from Reddit. The images have then been processed using OCR; in this way, I was able to convert the in-image text of each meme into a string. To identify consumer opinions, a topic extraction model has been developed based on these strings. In this GitHub page it is shown the whole Python code developed for this project.