In face of increasing competition from online retailers, traditional retailers have been losing their market share. One of the key disadvantages for traditional retailing, besides higher cost due to rental for shop space, labour for salesforce, etc., is the lack of data availability. Not like an online retailer, who can easily identify which customer is browsing what product and for how long, traditional retailers have very limited visibility in customers’ activities in the store. The absence of such valuable data has been the major obstacle for traditional retailers to better understand and optimise their retailing operations. In this project, I attempt to explore various sensors technologies that may lead to cost-effective ways of collecting data on customers’ activities in retailing outlets, investigate statistical methodologies for calibrating models with such data, and optimise operational decisions such as assortment, inventory, pricing, and product display. The framework closes the loop of data collection, parameter estimation, and model optimisation.