In 2016, I gave a presentation at PSU Elements after a year of data collection. The presentation was both about the methodology of data collection (below) as well as the conclusions, chiefly: the meaning of memes is a function of the meaning or context that usage has converged around, so there is, practically speaking, a "correct" way to use a meme, and using a meme "incorrectly" undermines that common understanding.
The methodology for data collection was via a nightly automated scraping process that scraped the "AdviceAnimals" subreddit (then receiving dozens of daily submissions of image macros), sanitizing the data and automatically inserting it into a database. This ran as a nightly job and emailed a status report.
There was a companion website (mostly defunct, presently) that used the data with some "games" that use the collected data, to help better understand the common understanding and intuitive understanding of different image macro templates.
(https://www.slideshare.net/AaronHill31/fantastic-memes-and-how-to-use-them-63058708">Slideshare, https://bit.ly/fantasticmemes">gSlides, https://github.com/armahillo/whichmeme">github)
The methodology for data collection was via a nightly automated scraping process that scraped the "AdviceAnimals" subreddit (then receiving dozens of daily submissions of image macros), sanitizing the data and automatically inserting it into a database. This ran as a nightly job and emailed a status report.
There was a companion website (mostly defunct, presently) that used the data with some "games" that use the collected data, to help better understand the common understanding and intuitive understanding of different image macro templates.
(https://www.slideshare.net/AaronHill31/fantastic-memes-and-how-to-use-them-63058708">Slideshare, https://bit.ly/fantasticmemes">gSlides, https://github.com/armahillo/whichmeme">github)