Everyday-Inspired Movies: Towards the Design of Movie Recommender Systems based on Everyday Life through Personal Social Media.

This paper proposes the idea of movie recommenders that draw from the users’ everyday life happenings as documented through their personal social media posts to produce relevant recommendations. We conducted an experimental study to understand the important dimensions to consider in the design of such a recommendation system. We began with the design hypothesis that matching keywords and categories from users’ social media posts to those from movie plots may increase users’ perceived relevance of movie recommendations. Our analysis revealed that beyond keywords and categories, emotional context and genre of movies are important aspects to consider. Based on these findings, we discuss the implications on the design of movie recommendation systems leveraging users’ everyday life through social media posts.