<p>Using store-level data from a major brick-and-mortar drugstore retail chain in Czechia, we outline and demonstrate a novel methodology that links the relative composition of sales in terms of basket types to location effects. Our research provides a major advantage over prior literature, where location effects on sales were considered separately from basket type classification. The new methodology put forward in this article can be used to analyze and optimize brick-and-mortar retailing. Benefits gained from this type of optimization apply both to the store operators and to their general neighborhood, as failing retailers and empty store areas bring negative externalities to residents (potential customers) and to nearby retailers. For the analysis presented, we combine location-related open access data sources with proprietary sales records of the retailer and introduce a two-step data-driven analysis method. First, sales/baskets are categorized into six distinct types: major baskets, large fill-in baskets, small fill-in, gifts &amp; treats, price specials, and diverse, budget-focused basket. Basket categorization is drawn from the features of individual purchases (financial value, number of products, basket diversity, and share of premium products). Second, categorized baskets are aggregated at the store level for analysis of location effects on sales. Regression models are used to assess the effects of prominent location-related factors (socio-demographic, proximity to amenities, street network data) on the relative composition of basket types. A detailed discussion is provided both in terms of the methodology applied and for the empirical results obtained.</p>