Abstract

Technology has made finding sex and companionship possible through online dating apps. The success of these online dating apps means that information on the dating market can be extracted. Using anonymized profile data scraped from Tinder in the Philippines, the study seeks to find differentiating characteristics between homosexual and heterosexual dating markets through clustering.

The feature used for clustering is the profile text data, which was vectorized using TF-IDF, and then clustered using Non-negative Matrix Factorization (NMF). Results show that while both groups have the same common interests, more homosexuals center their bios around their hobbies while more heterosexuals mention personal traits or preferences. Around twice as many homosexuals plug their social media on Tinder, inviting others to follow or message them there instead. This suggests that focusing campaigns on Instagram, or other social media platforms aside from Facebook, may be more effective when targeting the homosexual segment.