We introduce SemioGraph, that is, graphs whose vertices and edges are simultaneously mapped onto different systems of types or labels. To this end, we present a technique for visualizing SemioGraphs in an interactive manner. SemioGraph aims at coding as much information as possible within the same graph representation. This is interesting in cases such as word networks in which one has to visualize information units such as POS, node weight, node salience, node centrality etc. To showcase SemioGraph, we use word embedding networks. Word embeddings have become indispensable in the field of NLP, as they allow for significantly improving many tasks in machine learning. Therefore we built a website to facilitate the analysis of pre-trained word embedding models based on SemioGraph.