Daejeon, July 27–31
The suburban area of Beijing during the Republic of China (1912-1949) represents a critical transitional space in which imperial spatial legacies, Republican municipal planning, and new transportation technologies intersected. Yet this area has long been marginalized in historical geography research, which typically focuses on the walled city, major railways, or individual planning projects. This study adopts a Digital Humanities approach, integrating Historical Geographic Information Systems (HGIS), network analysis, and spatial statistics to reconstruct the transportation-settlement network of Beijing's suburbs. Based on the digitization of the 1947 Map of the Suburbs of Peiping, we model settlements as nodes, transport routes as weighted edges, and road grades as historically grounded indicators of infrastructural hierarchy. The research reveals a significant core-periphery differentiation: a highly developed High-High cluster in the western suburbs driven by cultural, educational, and tourist planning, contrasted with Low-Low clusters in the eastern and northern suburbs constrained by burial landscapes, canal decline, and weaker municipal investment. By correlating settlement density with transport-network density and node centrality, we identify modes of coordination, lag, and spatial mismatch, illustrating how Republican-era modernization selectively reshaped rather than uniformly integrated the urban fringe. This study contributes a new quantitative dataset for Beijing historical studies and situates suburban infrastructure within broader discussions of uneven urban development, historical GIS, and the transformation of twentieth-century Chinese cities.
Keywords: Digital Humanities; Historical GIS; Transportation Network; Spatial Analysis; Urban Periphery; Republican Beijing
Transportation networks serve as the arteries of urban-rural interaction, facilitating the flow of people, commodities, and information. Understanding the evolution of these networks is essential for revealing the historical processes of regional development. The suburbs of Beijing during the Republic of China (1912-1949) offer a unique case study. Unlike the rapid industrialization of port cities, Beijing's transformation was deeply rooted in its status as a former imperial capital. The suburbs were not merely agricultural hinterlands but complex spaces hosting emerging industries, modern education, military defenses, and leisure districts.
Scholarship on Beijing's transport history and urban morphology has long documented the formation of roads, administrative boundaries, and suburban functions in Chinese-language historical geography (Yin 2000; Sun / Xu 2013). At the same time, historical GIS has provided tools for reconstructing networks, measuring accessibility, and reflecting on how historical spatial evidence is transformed into analytical databases (Wang Chengjin et al. 2014; Chen et al. 2018; Gregory / Healey 2007; Harley 1989; Kitchin / Dodge 2007). English-language GIS work on Republican Beijing, especially by Billy K. L. So and collaborators, has demonstrated the value of spatial databases for studying urban institutions, services, and cultural patterns; however, that literature has focused mainly on the walled city and intra-urban social spaces rather than the transportation-settlement network of the suburban fringe (So 2010; So et al. 2012; Zhang et al. 2012; Wong et al. 2012). Building on these conversations, this study treats the 1947 Map of the Suburbs of Peiping not merely as a source to be visualized but as an official cartographic artifact whose road categories, settlement labels, and omissions require source criticism. It therefore asks a more explanatory question: did Republican-era infrastructural modernization integrate Beijing's suburbs, or did it selectively reinforce older imperial and economic spatial hierarchies? To answer this question, the paper constructs a historically weighted network database and examines three linked problems: the structural hierarchy of the network, its correlation with settlement distribution, and the historical forces that produced uneven suburban development.
The study focuses on the suburban area of Beijing, which originated from the "City-Affiliated" (Chengshu) jurisdiction of the Qing Dynasty (Han / Yin 1987; Hu 2016). Despite frequent administrative changes during the Republic, the geographic scope of this region remained remarkably stable until 1952 (Yin 2008). As Beijing underwent modernization, this area was formally integrated into the municipal administration and experienced a profound functional transformation, shifting from a traditional zone of agriculture and royal gardens to a diversified region hosting industry, education, and tourism (Sun / Wang 2016). This functional differentiation fostered a dense distribution of towns and villages, ranging from migrant settlements to Banner garrisons, connected by an increasingly complex transportation network. Distinct from the walled city yet integral to the capital's function, this suburban zone accumulated both traditional patterns and modern transitions. However, this unique and critical space has long been neglected by academia, necessitating a dedicated spatial analysis to understand its complex historical evolution.
The core spatial data for this study is derived from the Map of the Suburbs of Peiping (北平市城郊地图), surveyed and drawn by the Public Works Bureau of the Peiping Municipal Government in 1947 (Public Works Bureau of the Peiping Municipal Government 1947). As a high-precision official map (Scale 1:50,000), it provides a comprehensive snapshot of the topography, river systems, settlements, and transportation routes of the era.
To ensure historical accuracy and temporal depth, we cross-referenced this map with multi-source historical cartography and archival documents. Supplementary materials include the 1915 Map of the Four Suburbs of the Capital, the 1925 Map of Beijing and its Environs, and the 1934 Latest Detailed Map of Peiping (Map of the Four Suburbs of the Capital 1915; Map of Beijing and its Environs 1925; Latest Detailed Map of Peiping 1934). Additionally, textual archives such as Changes in Beijing's Transportation in Archives and Collection of Historical Materials on Urban Construction were utilized to verify road grades and construction timelines (Beijing Municipal Archives 2019; Chen Leren 2007). The earlier maps are used here as reference points rather than as a full diachronic database: they help distinguish inherited arteries from routes that were upgraded or emphasized by Republican municipal planning, while a complete multi-temporal network reconstruction remains a future task.
Network construction proceeded in four steps. First, settlements, towns, villages, railway stations, and named suburban nodes shown on the historical maps were digitized as nodes. Second, railways and roads were digitized as edges when the map indicated a direct route between nodes; intersections and route junctions were used to preserve topological connectivity, but the abstraction does not claim to capture actual travel volume. Third, all edges were assigned a road category from the map legend and archival verification: railways, asphalt roads, crushed stone roads, earth roads, and country paths (see Figure 1). Fourth, the node-edge table was exported from ArcGIS Pro to Gephi for network calculation and then rejoined to the GIS database for spatial analysis.
To quantify relative infrastructural importance, we applied a weighting system guided by two principles: the qualitative hierarchy defined by Republican official engineering standards (1929/1934) (Chinese Institute of Engineers 1929; Hubei Provincial Government 1934) and the quantitative Rank Sum (RS) method (Chergui / Jiménez-Martín 2024; Roszkowska 2013). For non-specialist readers, the Rank Sum approach is used here in its ordinal sense: because the sources rank transport facilities from higher to lower technical standard, we convert that historical order into a simple descending score. Accordingly, weights were assigned as follows: Railways (5), Asphalt Roads (4), Crushed Stone Roads (3), Earth Roads (2), and Country Paths (1). Weighted degree centrality is calculated as the sum of the weights of all routes directly connected to a settlement. The metric therefore measures potential local accessibility and infrastructural advantage, not the actual number of travelers or goods moving through the node. These metrics were then reintegrated into GIS to analyze spatial patterns and correlations.
Figure 1. The digitized Map of the Suburbs of Peiping and the GIS database structure.
The transportation network of Republican Beijing evolved on the foundation of historical thoroughfares, forming a distinct hierarchical structure. By calculating the weighted degree centrality based on road weights and connections (Figure 2a), we identified a sharp differentiation in node importance. High-centrality settlements (marked in red), such as Haidian, Fengtai, Wanping, and the areas outside Xizhimen and Yongdingmen, occupy the core positions, often serving as critical hubs associated with railway stations and possessing strong distribution capabilities. Secondary nodes (yellow) exhibit a "string of pearls" pattern radiating from city gates, a trend most pronounced in the Western and Southern suburbs.
Specifically, routes from Xizhimen and Fuchengmen extend northwest towards the "Three Hills and Five Gardens" district, while the Southern routes connect to Nanyuan and Fengtai, creating strong radial axes. In contrast, the Eastern and Northern suburbs display a significant lag in development; settlements here (blue and grey) are sparsely distributed and connected mainly by low-grade earth roads or country paths, indicating weak connectivity with the urban core. This unbalanced spatial pattern is further confirmed by the Kernel Density Estimation (KDE) (Figure 2b), which highlights two high-density belts: a "North-South" strip in the West and an "East-West" strip in the South, while the vast eastern and northern peripheries remain marginalized as low-density zones.
Figure 2. (a) The weighted degree centrality of settlements; (b) Kernel Density Estimation (KDE) of settlement centrality.
Using Local Spatial Autocorrelation (LISA), we visualized significant node-level clustering patterns (Figure 3), revealing a stark core-periphery differentiation. A prominent High-High (HH) cluster (Hot Spot) dominates the Western Suburbs, forming a dense block around the "Three Hills and Five Gardens" district and extending along major arteries. This indicates a zone of synergistic development where accessible towns were surrounded by other highly accessible settlements.
Conversely, extensive Low-Low (LL) clusters (Cold Spots) characterize the Eastern and Northern Suburbs. The East Suburb features clustered cold spots near Guangqumen and boundary areas like Guanzhuang, while the North Suburb shows a dispersed, homogenous low-value distribution. These areas suffered from sparse nodes and delayed infrastructure construction.
Crucially, the analysis identified spatial outliers. Low-High (LH) outliers appear on the fringes of the Western cluster, reflecting a "shadow effect" where peripheral villages failed to integrate despite proximity to major hubs. Meanwhile, scattered High-Low (HL) outliers in the remote suburbs, such as Qinghe and Gaobeidian, validate Central Place Theory: despite being in underdeveloped regions, these towns evolved into local sub-centers due to their administrative or strategic transport roles.
Figure 3. (a) Local Spatial Autocorrelation (LISA) clustering map; (b) Cold and Hot Spot analysis.
To examine the interaction between settlement patterns and network development, we conducted a bivariate correlation analysis. The results show a strong positive correlation (R²=0.64) between settlement density and transport network density, but the spatial distribution is heterogeneous. We identified three distinct modes. First, the Coordination Mode (High-High) is primarily found in the Western and Southern suburbs (e.g., outside Yongdingmen), where dense settlements overlap with developed infrastructure, creating a virtuous cycle. Second, the Lagging Mode (Low-Low) is prevalent in the vast Eastern fringe, where both settlement and transport density remain low due to a lack of development focus. Third, a Spatial Mismatch is observed along western mining railways. In these corridors, transport density is high due to rail lines, but settlement density is low, creating a "corridor effect" where infrastructure served resource extraction rather than local community integration (Figure 4).
Figure 4. Bivariate map showing the spatial correlation between settlement kernel density and transportation network density.
Based on historical context and spatial patterns, we identify three key driving forces. They should be understood as mechanisms of selective modernization: new transport facilities did not simply replace older spatial orders but interacted with imperial landscapes, municipal priorities, and strategic infrastructures to produce uneven access.
Legacies of Imperial Power. The spatial differentiation was fundamentally rooted in the Qing Dynasty's layout. The Western Suburbs benefited from the infrastructure of Royal Gardens and Banner garrisons, while the Eastern Suburbs were restricted by the presence of Royal Mausoleums and the decline of the Grand Canal, leading to long-term underdevelopment.
Urban Planning Policies. The Republican government reinforced this disparity through selective modernization. Plans such as the Construction Plan for Tourist Districts (1933) explicitly targeted the Western Suburbs for cultural and tourism development, channeling resources into upgrading roads (asphalt/stone) in these areas while neglecting others (Peiping Municipal Government 1934).
Modern Infrastructure Construction. Railways and highways changed the meaning of distance in the suburbs, but their impact was selective. Lines such as the Peking-Kalgan Railway and mining corridors improved regional connectivity and resource extraction capacity, while often bypassing nearby village networks. The result was a corridor effect: modern infrastructure could intensify specific strategic or economic functions without necessarily integrating local communities into a balanced suburban network.
This study investigates transportation infrastructure and settlement patterns in the suburbs of Republican Beijing as a problem of uneven urban development. The results confirm a core-periphery structure in which the Western and Southern suburbs emerged as concentrated hubs of accessibility, while eastern and northern fringe areas remained weakly connected. The key historical implication is that Republican-era modernization did not uniformly integrate the urban fringe; rather, it selectively amplified inherited spatial advantages around gardens, gates, railways, and planned tourist or educational zones.
Methodologically, the contribution of this paper lies less in inventing new GIS tools than in building a transparent historical GIS workflow: an official 1947 map is converted into a weighted transportation-settlement network, the weighting assumptions are grounded in Republican engineering standards, and the resulting metrics are interpreted alongside spatial statistics and historical context. This workflow also requires critical cartographic awareness. The map's categories are products of municipal surveying and governance, so digitization transforms a historically situated representation into analytical data rather than recovering a neutral landscape. By foregrounding this process, the paper extends existing GIS studies of Republican Beijing from intra-urban institutions toward suburban infrastructure and shows how quantitative spatial analysis can refine, rather than replace, historical interpretation.
The study also has clear limitations. It remains primarily a 1947 cross-sectional analysis, although the 1915, 1925, and 1934 maps are used to contextualize continuity and change. Future work should construct a multi-temporal network database to compare route addition, road upgrading, and settlement change across several benchmark years. The current analysis also focuses on physical topology and potential accessibility; due to data constraints, actual flows of people, goods, military traffic, and coal transport remain only partially visible. Integrating socio-economic records, traffic data, and archival evidence will allow future research to test more directly how transport infrastructure reshaped urban-rural relations.