This paper explores the emerging paradigm of alternative food networks (AFN), with a focus on AMAP (Association pour le Maintien de l'Agriculture Paysanne) in France as a case study. Delving into the multifaceted concept of spatial proximity within such networks, this paper explores three hypotheses drawn from literature on the spatiality of AFN, analyzing both 'local' distances variability and the characteristics of agricultural and urban contexts where such local-based AFN take place. Methodologically, the study leverages an original, manually compiled database at a national scale encompassing AMAP producers and distribution points. Specific analytical protocols are developed combining traditional geographical approaches with machine learning techniques. Key findings reveal the influence of both population density and agricultural land availability on the distances between producers and selling locations. Moreover, the study discerns that the nature of products and their processing levels significantly shapes 'local' distances. Additionally, the paper offers insights into distinctive features of the morphological landscape associated with AMAP producers. These findings may serve as a catalyst for future inquiries into the spatial dynamics and potential spatial configurations of alternative food networks.