Four real-world user-item bipartite networks, i.e., Movielens, Rate Your Music (RYM), Epinions and Douban, are used in this paper to explore heterogeneous interactions in online commercial networks. The raw data used in this paper include rating information with integer rating scales to one item by a user. This type of rating data are extracted from several major websites where users are allowed to express their opinion about certain online products such as movies, music and books.