Research and Application of Data-driven Cigarette Retailer Classification: A Grid-Based Approach

In view of the deficiencies in the current cigarette retailer grade division method in accurately reflecting the business operation capabilities of retailers, this study innovatively puts forward a hierarchical classification method driven by big data. This method divides the cigarette sales regional market into business circle grids of the same size. Subsequently, it conducts systematic preprocessing on external environment data such as consumption level, population density and Points of Interest (POI) within the grids. On this basis, it carries out hierarchical and classified management of cigarette retailers to achieve the adaptation between cigarette sales channels and different business circle environments and improve the accuracy of cigarette supply allocation. Taking the Yulin market in Guangxi as an example, the business circle grids in the Yulin market are divided into 5 special business circles and 7 non-special business circles, which increases the combination of goods supply from the original 30 to 360, increases the subdivision degree of the supply area by 1100%, and raises the utilization rate of the goods of the new Zhenlong brand cigarettes from 35% to 59%.