Where to place the next outlet? Harnessing cross-space urban data for multi-scale chain store recommendation

Chain store has become an important business form in modern society. For the same chain group, they often have inner store classification regarding the store scale to meet the service requests and profit optimization needs of different areas. In this paper, we present ChainRec, a framework for chain store placement recommendation considering its scale. Specifically, we extract three types of associative features from crossspace data sources, including geographic features, commercial features, as well as scale features. Based on these features, we adopt supervised regression and classification to solve two scale-specific chain store placement problems. Experiments with online and offline datasets from the Chengdu City in China validate the effectiveness of our framework.