Questioning the algorithmic transparency of location-based platforms

In the platform society algorithms are perceived as ‘black boxes’ (Pasquale, 2015) and users have only a vague understanding of the criteria they adopt to select information. Location-based platforms algorithms influence the visibility of different points of interest (POI), thus shaping the user interaction with venues and places.    The paper adapts the Diakopoulos and Koliska model (2017) and presents a new framework for analyzing the algorithmic transparency of location-based platforms. Research questions are the following: RQ1) How location-based platforms communicate algorithmic transparency?; RQ2) Which are the most relevant dimensions they take into consideration (data, model, inference, interface)?; RQ3) How platforms communicate transparency toward different targets (i.e. consumers and suppliers)? Following Rader, Cotter and Cho (2018) we expect location-based platforms are less transparent about the data they manage and about their model they use and more transparent about the inferences. Moreover, we expect location-based platforms are more transparent toward suppliers rather than consumers.   The paper analyzes how 3 very popular location-based platforms (Google Maps, Tripadvisor and Instagram) disclose algorithmic transparency as it emerges from the analysis of ‘extant’ online data officially released (policies, guidelines, and tutorials) and from the analysis of the platforms’ mobile interface. The analysis revealed platforms are less transparent about the data they manage and model they use, and more transparent, only toward suppliers, about the inferences they propose. Moreover, location-based platforms are more transparent toward suppliers rather than consumers; indeed, commercial interests favours the algorithmic transparency and visibility of location-based content.