Platform Urbanism and “Splintering Amenitization”: An Analysis of Canadian Cities

Abstract

This article engages with the spatialities of platform urbanism by foregrounding where digital platforms are located in cities. Drawing on a geocoded data set of visible, material traces of platformization collected across neighborhoods in Toronto, Vancouver, and Montreal, we consider the influences that characteristics of urban built environments—including existing amenities, urban morphology, and area-level socioeconomic factors—have on platforms’ locations. Through a Poisson regression of these variables, we find that the presence of existing urban amenities most strongly explains the locations of material traces of urban platformization on the cityscape at the city block scale. We position platforms themselves as a novel amenity class that extends emplaced utility and lifestyle functions to urban residents. In so doing, we contend that the platformization of urban landscapes constitutes a form of “splintering amenitization,” wherein platformized urban amenities demonstrate spatial patterns of colocating with other, existing urban amenities in already amenity-rich areas to the exclusion of amenity-poor enclaves. This, we argue, is important because neighborhoods’ abilities to attract amenities are central to how enclaves both position themselves and compete for status within urban spatial hierarchies.

通过强调数字平台在城市中的位置, 本文探讨了平台城市主义的空间性。根据多伦多、温哥华和蒙特利尔社区平台化的可见物质痕迹地理数据, 我们考虑了城市建筑环境特征(现有便利设施、城市形态和局地社会经济因素)对平台位置的影响。根据对这些变量的泊松回归分析, 我们发现, 城市现有便利设施能最有力地解释街区尺度的城市平台化物质痕迹在城市景观中的位置。我们将平台定位为一种新的便利设施, 能将城市的功能性和生活性功能扩展到城市居民。我们认为, 城市景观的平台化构成了“碎片化的便利设施化”。在便利设施丰富地区, 平台化的城市便利设施与其它现有城市便利设施在空间分布上相吻合, 但是便利设施较差的飞地被排除在外。这一点很重要, 社区吸引便利设施的能力对于在城市空间层次结构中对飞地进行定位、开展竞争, 具有核心作用。

En este artículo se abordan las espacialidades del urbanismo de plataformas, poniendo en primer plano la cuestión de dónde son ubicadas las plataformas digitales en las ciudades. A partir de un conjunto de datos geocodificados sobre rastros materiales visibles de la plataformalización, recopilados en vecindarios de Toronto, Vancouver y Montreal, consideramos las influencias que tienen las características de los entornos urbanos edificados–incluidas las comodidades existentes, la morfología urbana y los factores socioeconómicos a nivel de área–sobre la localización de las plataformas. Por medio de una regresión de Poisson de estas variables, descubrimos que la presencia de comodidades urbanas existentes explica con más fuerza las ubicaciones de rastros materiales de la plataformalización urbana en el paisaje urbano, a la escala de manzana. Posicionamos las plataformas como una nueva clase de servicio en sí mismas que extiende las existentes funciones de utilidad y estilo de vida de los urbanitas. Al hacerlo, sostenemos que la plataformalización de los paisajes urbanos constituye una forma de “amenitización astillada” en la cual las comodidades urbanas plataformalizadas se demuestran los patrones urbanos de colocación con otras comodidades urbanas existentes en las áreas ya enriquecidas con comodidades, excluyendo los enclaves pobres en comodidades. Esto, argüimos nosotros, es muy importante porque la capacidad de los vecindarios para atraer nuevos servicios es fundamental para determinar cómo se posicionan a sí mismo los enclaves y compiten por estatus dentro de las jerarquías espaciales urbanas.

Key Words:

• amenities
• cities
• platforms
• regression analysis
• splintering urbanism

关键词:

• 便利设施
• 城市
• 平台
• 回归分析
• 碎片化城市主义

Palabras clave:

• análisis de regresión
• ciudades
• comodidades
• plataformas
• urbanismo fragmentado

Disclosure Statement

No potential conflict of interest was reported by the authors.

Supplemental Material

Supplemental data for this article can be accessed on the publisher’s site at: https://doi.org/10.1080/24694452.2024.2322474

Notes

1 Census tracts are units of Canadian census geography that exist in metropolitan areas. They are defined as “small, relatively stable geographic areas that usually have a population of less than 10,000 persons” (Statistics Canada 2016a). Census tracts are accepted proxies for neighborhoods in the Canadian urban context (Walks 2013).

2 The following typologies were excluded: “corridor,” “footway,” “living_street,” “path,” “raceway,” “steps,” and “tracks.”

3 400 m corresponds to “the theoretical minimum of a pedshed, or walkable catchment” equivalent to five minutes of walking, which “is generally considered acceptable in the urban areas of car-oriented places such as … North America[n cities]” (Mateo-Babiano et al. 2016, 298; see also Boone et al. 2009).

4 As per the Employment Equity Act of Canada, the term visible minority refers to “persons, other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour” (Statistics Canada 2017).

5 A dissemination area is a unit of Canadian census geography finer than a census tract. It typically spans multiple street blocks and contains between 400 and 700 residents (Statistics Canada 2016b). Street blocks in the same DA are all assigned the same values for the socioeconomic variables. This potentially introduces some error in the data, but unfortunately these data are not available at a finer spatial resolution.

6 Please refer to Table 2 for an explanation of each term.

7 Because the Poisson regression uses a log-link specification, the coefficients, which normally indicate the expected difference in the logs of expected platform counts, can be exponentiated to infer a relative increase. In this case, a coefficient of −0.022 means a 1-unit increase in the percentage of visible minorities and corresponds to a exp(−0.022) ∼ 0.98 factor, or a 2 percent decrease.

8 We should note here that the residuals do not exhibit spatial autocorrelation (Moran’s I = 0.05, p = 4.759 × 10⁻⁷). To account for potential overdispersion or zero-inflation in the model (many street blocks have no platform presence), we also ran a version of the same model with quasi-Poisson, negative binomial and zero-inflated Poisson specification. Similarly, we also included a specification that splits variables between the value within the street block itself and the 400-m buffer (to separate the immediate and nearby spatial context). The results of these models are included in Supplemental Tables S.1 and S.2. Because the results are not sensitive to these alternative specifications and are consistent with the results discussed so far, we only report the Poisson model here.

Additional information

Funding

Anirudh Govind’s contributions to this article were financially supported by patronage funding provided to KU Leuven for carrying out fundamental scientific research into urban change. Agnieszka Leszczynski’s contributions to this article were supported by the Social Sciences and Humanities Research Council of Canada (Grant No. 430-2019-00233). Ate Poorthuis’s contributions to this article were supported by Internal Funds KU Leuven Grant Number STG/20/021.

Notes on contributors

Anirudh Govind

ANIRUDH GOVIND is a Doctoral Researcher at the Public Governance Institute and the Department of Earth and Environmental Sciences at KU Leuven, Leuven, Belgium. E-mail: [email protected]. His research examines the built environment’s influences on social interactions and sociospatial processes.

Agnieszka Leszczynski

AGNIESZKA LESZCZYNSKI is Associate Professor in Geography and Environment at Western University, London, ON N6A 3K7, Canada. E-mail: [email protected]. Her current work focuses on cities, platforms, and aesthetics.

Ate Poorthuis

ATE POORTHUIS is Assistant Professor in the Department of Earth and Environmental Sciences at KU Leuven, Leuven, Belgium. E-mail: [email protected]. His work uses spatial analysis, big data, and visualization to analyze the intersection of human mobility and spatial inequality in urban systems.