During recent years, the amount of multimedia data on social websites is growing exponentially. It is observed that multimedia data corresponding to the same semantic concept usually appears in different media types and from heterogeneous data sources. In order to synchronize and leverage these diverse forms of media data for multimedia applications, we present a real-world web dataset collected from Google, Flickr and YouTube for cross-media research. The dataset includes 41,387 text files, 65,371 images and 30,818 videos (about 1091 hours) which are correlated semantically with each other by 335 representative visual concepts. Widely-used features are extracted for each media type and all of them are publicly available. To evaluate the performance of our dataset, experiments on baseline recognition, feature evaluation and domain adaptation are performed. The experimental results indicate that it is possible to perform multiple cross-media tasks based on our proposed dataset.