Retailer Service Point Placement via Customer Geo-Location Clustering

Determining the location of a physical service point (e.g., store branch) is of utmost importance for the success of a retailer company that relies on a distributed branch network to promote its goods and services. The distance from the nearest service point is an important factor to consider, as it is a measure of convenience for its customers. The initial set-up costs for the establishment of a new service point should be carefully weighed against the expected benefits, as future relocations or new establishments translate to additional, unnecessary costs. In this Data Science study, we extract data from the Cyprus National Open Data portal in reference to Cyprus-based legal entities to determine both the location and size of the service points of a retailer serving the population. To achieve this, we model the challenge as a maximal coverage location problem and use unsupervised clustering to develop our solution. The results of our machine learning model are the optimal number and geographical positions for the service points of the retailer, based on the actual location of the points of the retailer’s customers. We statistically evaluate our results by comparing them with the service point placement of two competitor companies. We show that the service points proposed by our model minimize the distance traveled by customers to the nearest service point, while also utilizing fewer service points and resulting in better customer service with reduced operating costs.