Multivariate Features Based Instagram Post Analysis to Enrich User Experience

Abstract

In today’s digital world, wherein user personalized content such as text, video, photos and much more have become an integral part of people’s daily lives; photo intensive social media applications have acquired enhanced adoption in social media users through Instagram. By the time, Instagram- a photo sharing site continues to evolve and grow in popularity. Although Instagram has rapidly gained popularity among active social media users, analysis of the interaction and engagement among people on Instagram is missing and is almost an untouched area. In this paper, we inspect some prominent user interaction properties and photo properties to understand users’ engagements towards posts on Instagram. The considered user interaction properties are hashtags, photo post time etc as users’ posted photo context. On the other end, photo properties are user’s posted photo features or image contexts such as image filters. We have performed these user interaction properties and photo properties analysis task on eight major cities’ Instagram posts and further classified the posts of these eight cities in five categories using Non-negative matrix factorization and latent Dirichlet allocation algorithm. The four prime influencing analyses have been computed to get ecology of the users on Instagram photo posts, which are Time based analysis (TBA), Image Filter analysis (IFA), Image Hashtags analysis (IHA) and Image categorization analysis (ICA). Henceforth, this multivariate feature based Instagram analysis will help users to gain insight of popular content and make their respective content popular so as to reach out to a maximum number of people.