Economy and Geography Shape the Collective Attention of Cities

Ke-ke Shang, Jiangli Zhu, Junfan Yi, Liwen Zhang, Junjie Yang, Ge Guo, Zixuan Jin, Michael Small · 2025

Read Economy and Geography Shape the Collective Attention of Cities by Ke-ke Shang, Jiangli Zhu, Junfan Yi, Liwen Zhang, Junjie Yang, Ge Guo, Zixuan Jin, Michael Small (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-08907.txt.

Introduction

Ke-ke Shang and co-authors study what draws a city's collective attention online in a 2025 paper. They build attention networks from time-series data for the United States and China, then ask how wealth and geography shape the patterns at several scales. Past network studies of human behavior had mostly ignored those two forces.

At the broad scale, a city's attention tracked closely with its economic output. Inside network communities, rich cities sat at the core and held handy geographic spots. They also served as bridges between communities, with influence in line with their output. Tightly linked cities tended to be near each other and share topics. One pattern split the two countries: population size predicted attention in the United States but not in China. A combined index of people, access, and scenery still failed to explain the gaps.

Catalog record

CREATORKe-ke Shang, Jiangli Zhu, Junfan Yi, Liwen Zhang, Junjie Yang, Ge Guo, Zixuan Jin, Michael Small
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-08907

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Economy and Geography Shape the Collective Attention of Cities Ke-ke Shang, Jiangli Zhu, Junfan Yi, Liwen Zhang, Junjie Yang, Ge Guo, Zixuan Jin, Michael Small Abstract Complex networks are commonly used to explore human behavior. However, previous studies largely overlooked the geographical and economic factors embedded in collective attention. To address this, we construct attention networks from time-series data for the United States and China, each a key economic power in the West and the East, respectively. We reveal a strong macroscale correlation between urban attention and Gross Domestic Product (GDP). At the mesoscale, community detection of attention networks shows that high-GDP cities consistently act as core nodes within their communities and occupy strategic geographic positions. At the microscale, structural hole theory identifies these cities as key connectors between communities, with influence proportional to economic output. Overlapping community detection further reveals tightly connected urban clusters, prompting us to introduce geographic and topic-based metrics, which show that closely linked cities are spatially proximate and topically coherent. Of course, not all patterns were consistent across regions. A notable distinction emerged in the relationship between population size and urban attention, which was evident in the United States but absent in China. Building on these insights, we integrate key variables reflecting GDP, geography, and scenic resources into regression model to cross-verify the influence of economic and geographic factors on collective user attention, and unexpectedly discover that a composite index of population, access, and scenery fails to account for cross-city variations in attention. Our study bridges the gap between economic prosperity and geographic centrality in shaping urban attention landscapes. keywords Attention networks, community detection, structural hole, economic prosperity, geographic centrality † † equal-contributors: These authors contributed equally to this work. † † equal-contributors: These authors contributed equally to this work. † † equal-contributors: These authors contributed equally to this work. Herbert Alexander Simon introduces the concept of attention economics, highlighting that with growing information abundance, human attention constitutes a scarce resource simon1973applying . This foundational insight has given rise to the study of collective attention, where large groups of users simultaneously focus on specific digital content, revealing competitive dynamics that shape online engagement and information diffusion wu2007novelty ; cattuto2009collective . On the other hand, social scientists have increasingly acknowledged the potential and significance of applying complex network theory in urban planning, which is utilized for traffic flow prediction and population pattern recognition wang2024role . With the groundbreaking achievements of complex networks barabasi1999emergence ; barabasi2005origin , an emerging research trend seeks to adopt attention networks, which were originally developed to model the dynamics of collective attention wu2007novelty ; cattuto2009collective ; lehmann2012dynamical , to elucidate issues related to urban development and human activities Eubank2004ModellingDO ; kontokosta2024socio , addressing the complexity and uncertainty inherent in urban systems. The fundamental role of attention networks in understanding how urban identities are constructed and internalized within collective human cognition is increasingly recognized. However, while current research primarily revolves around the impact of attention networks on urban morphology and mechanisms zhao2024unravelling , the cognitive and cultural foundations of the urban attention component within these networks remain underexplored. This gap calls for a focus on the very nature of the city genome embedded in public perception. Moreover, community detection focuses on identifying cohesive groups within complex networks girvan2002community . Communities, fundamental substructures in network science, are formally defined as groups with significantly denser internal edges than external ones shang2020novel . Rooted in the tenets of community detection, in the interconnected digital realm, recent studies observe the dynamics of collective attention, particularly in the tourism domain, intricately intertwined with the geographical and economic factors of cities xu2022tourism ; kim2020exploring . Following those perspectives, we mainly explore attention networks utilizing community detection, complemented by conventional economic methodologies for cross-validation. We construct attention networks using dynamic shifts in public focus during major lockdown periods in the United States and China, two leading economies with distinct cultural contexts, when restricted movement drove digital engagement and the easing of restrictions triggered a sharp rise in travel-related attention (Extended Data Figure 4 ), revealing an intrinsic and deeply rooted collective attention toward urban destinations. Specifically, user-generated content and behavioral data from TripAdvisor and Ctrip serve as high-fidelity proxies for collective attention, capturing public interest in the United States and China, respectively. Each node in those networks represents a city, and a link between two nodes is established when users mention both cities in their travel itineraries, indicating the flow of user attention. The emergent complex system, formed by the aggregation of all such user-generated connections, reflects the dynamics of collective attention. The attention network in the United States comprises 396 nodes and 1865 edges, and in China, it consists of 128 nodes and 461 edges (Extended Data Figure 5 ). To strengthen our analysis and bridge the domains of network science and computational economics, we incorporated comprehensive economic and geographical data from various official sources. Economic and demographic statistics were sourced from the United States Bureau of Economic Analysis, and the National Bureau of Statistics of China. Geographical features, including administrative boundaries, railway stations, and airports, were obtained from 12306 China Railway, OpenStreetMap, along with official government repositories (Extended Data Table 3 ). 1 Uncovering GDP footprint in collective attention In assessing online social attention, the volume of referenced content stands as a natural metric for collective attention. From a macro-perspective, we examine the attention rankings of various cities ( Figure 1 ) and find that economically dominant metropolitan areas consistently rank among the top ten in collective user attention. Notable examples from the United States include New York County, which encompasses the economic core of Manhattan, and Clark County, anchored by Las Vegas. In China, Beijing serves as the political center and is also one of the most economically powerful cities, while Shanghai stands out as a major financial hub. Figure 1: The ranking changes of cities in the top ten for consecutive four months. Each time a city is mentioned by a user post, it counted as one mention. The cities are then ranked based on their monthly mention counts. Across the United States and China, a consistent presence of seven cities each has been observed in the rankings. To ensure a comparable network scale, we use counties in the United States and prefecture-level cities in China as the geographic units of analysis. For consistency, we refer to both types of administrative units as cities. All cities analyzed in this paper are located within the contiguous United States and mainland China. Given that city classifications typically incorporate both economic output and population size as key criteria, we further investigate the relative influence of these factors on online visibility. To evaluate the independent contributions of GDP and population to the number of city mentions, we conduct a multiple linear regression analysis (LR) galton1886regression . This model allows us to assess how each factor, GDP and population, affects city mentions while controlling for the other. Following standard econometric practice, T-tests were applied to assess the statistical significance of the coefficients. As shown in Table 1 , the results reveal a robust positive association between GDP and city mention counts across major urban centers in both countries. However, the population

Read the complete text