Music and singers are influential in local society. An in-depth study on singers is beneficial to various sectors. However, the evolutional characteristic and the daunting complexity of the interrelationship among singers made the problem technically intriguing. In this paper, we present a novel commentary-based social network analysis (CBSNA) methodology to analyze the singer relationships. Developing weighting schemes and adopting k-nearest-neighbors (kNN) approach from network theory as a visualization technique, we simplify the resulting dense network to ease understanding and further investigations. Proof-of-concept experiments are conducted by using two popular datasets to verify the effectiveness of the proposed approach and the empirical results are promising.