Online social networking services, such as Twitter and Facebook have attracted considerable research interests. Event detection has been studied for quite some time, and there are studies that discuss event detection on Twitter; social network analysis has been studied for an even longer time, and there are studies that apply social network analysis to data collected from Facebook. However, not much research attention is on event detection on Facebook. In this paper, we address the problem of how to detect events in an ego network on Facebook. Our proposed approach first uses K-Means to cluster posts based on words, then builds an interaction graph based on comments and likes given to posts, then applies PageRank to the interaction graph in order to identify active posters, and finally finds the topics based on the frequent words used by the active posters. Based on the experiment result, our proposed approach can identify topics that are highly relevant to real-world events and simultaneously identify users who are of higher degrees of interaction.