Popular node embedding methods such as DeepWalk follow the paradigm of performing random walks on the graph, and then requiring each node to be proximate to those appearing along with it. Though proved to be successful in various tasks, this paradigm reduces a graph with topology to a set of sequential sentences, thus omitting global information. To produce global-aware node embeddings, we propose BiasedWalk, a biased random walk strategy that favors nodes with similar semantics. Empirical evidence suggests BiasedWalk can generally enhance global awareness of the generated embeddings.