The movements of individuals within and among cities influence critical aspects of our society, such as well-being, the spreading of epidemics, and the quality of the environment. When information about mobility flows is not available for a particular region of interest, we must rely on mathematical models to generate them. In this work, we propose Deep Gravity, an effective model to generate flow probabilities that exploits many features (e.g., land use, road network, transport, food, health facilities) extracted from voluntary geographic data, and uses deep neural networks to discover non-linear relationships between those features and mobility flows. Our experiments, conducted on mobility flows in England, Italy, and New York State, show that Deep Gravity achieves a significant increase in performance, especially in densely populated regions of interest, with respect to the classic gravity model and models that do not use deep neural networks or geographic data. Deep Gravity has good generalization capability, generating realistic flows also for geographic areas for which there is no data availability for training. Finally, we show how flows generated by Deep Gravity may be explained in terms of the geographic features and highlight crucial differences among the three considered countries interpreting the model's prediction with explainable AI techniques.© 2021. The Author(s).

The ongoing coronavirus disease 2019 (COVID-19) outbreak expanded rapidly throughout China. Major behavioral, clinical, and state interventions were undertaken to mitigate the epidemic and prevent the persistence of the virus in human populations in China and worldwide. It remains unclear how these unprecedented interventions, including travel restrictions, affected COVID-19 spread in China. We used real-time mobility data from Wuhan and detailed case data including travel history to elucidate the role of case importation in transmission in cities across China and to ascertain the impact of control measures. Early on, the spatial distribution of COVID-19 cases in China was explained well by human mobility data. After the implementation of control measures, this correlation dropped and growth rates became negative in most locations, although shifts in the demographics of reported cases were still indicative of local chains of transmission outside of Wuhan. This study shows that the drastic control measures implemented in China substantially mitigated the spread of COVID-19.Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works.

空间交互模型被广泛应用于地理要素关系强度的模拟,然而目前大量研究或建立在模型参数标定理想化、模式化的假设条件下,或是在暗箱中完成,由此导致模拟结果与实际的偏差却被严重低估。基于2015年中国春运期间人口省际流动的城市间O-D数据,在逐日、分市的研究精度下,实证推算人口流动重力模型变量的回归系数,探究模型代理变量影响效应的空间异质性,并评估重力模型在人口流动模拟上的误差。结果显示：① 重力模型参数标定的复杂性体现在交互对象代理变量影响程度的非对称性,和变量回归系数的空间异质性随研究精度加深显著加剧两个方面,因此模型参数标定的模式化将导致估算结果空间差异的趋势收敛;② 2015年春运期间中国人口省际流动距离衰减系数为1.970,在地级行政单元视角下,人口流出地距离衰减系数值域为0.712（驻马店）~7.699（乌鲁木齐）,人口流入地系数值域为0.792（三亚）~8.223（乌鲁木齐）;③ 应用重力模型模拟人口流动结果与实测流（百度迁徙数据）存在显著误差。就加权绝对平均误差而言,拟合总误差为85.54%,其中空间相互作用效应造成了86.09%的实测流与模拟流的最大误差,相对流出力、相对吸引力分别造成57.73%、49.34%的模型误差。因此,空间交互效应仍然是当前最难以模式化的因素。

Simulations based on spatial interaction models have been widely applied to understand the strength of relationships between geographical elements, but many issues remain unclear and deviations between actual and simulated results have often been seriously underestimated. A high-precision Baidu migration process combined with mass relationships is applied in this study and enables the generation of regression coefficients of gravity model based on programmed large-scale regression simulations. A series of accuracy assessments are then developed for 2015 empirical projection daily regression coefficients that can be applied to Chinese spring interprovincial mobile gravity model variables as well as spatiotemporal research that utilizes regression coefficients within a heterogeneity research model. This approach also enables the error within the gravity model to be assessed in terms of floating population simulations. The results of this analysis lead to a number of clear conclusions, including the fact that parameter calibration complexity for the Chinese population mobility gravity model is reflected in the degree of influence asymmetry within spatial object interaction variables, and that the spatial heterogeneity of the variable regression coefficient increases in two distinct fashions. The first of these increases has to do with the overall influence of specific variables, including the fact that differences between proxies tend to be higher than inflow-outflow characteristics. In contrast, the second set of increases is related to economic levels, industrial scales, the proportion of the tertiary industry, and public service facilities. In this latter case, two-way population flow exerts a more profound influence on results and thus the scope of possible explanations for phenomena is more extensive. The regression coefficient for the existence of positive and negative proxy variables therefore relates to differences in spatial heterogeneity, including at the city level, and also assumes that floating population gravity model regression coefficients ignore spatiotemporal changes in the heterogeneity coefficient. This leads to spatial differences in estimated results and thus convergence trends, but further enables the identification of anisotropic interactions in extension space. The second main conclusion of this research is that the national scale population flow distance attenuation coefficient was 1.970 during the spring of 2015, while at the level of prefectural administrative units and given population outflow, the range encapsulated by this coefficient fell between 0.712 (Zhumadian) and 7.699 (Urumqi). Data also reveal a population inflow coefficient of 0.792 for this year that ranged as high as 8.223 in both Sanya and Urumqi. Population flow simulation results using the gravity model and including Baidu migration measured flow data were also subject to significant error. Third, the results of this analysis reveal a total fitting error of 85.54% in weighted absolute mean; the spatial interaction effect within this is responsible for a maximum error of 86.09% in actual and simulated flows, while relative outflow force and attractiveness encompass 57.73% and 49.34% of model error, respectively. These results show that the spatial interaction effect remains most difficult to model in terms of current factors.

During the past 50 years, profound changes took place in the field of spatial structure of Chinese urban system. This paper will focus on the spatial relations (including network relations and territorial relations) in the inner urban system, and classify the ranks of spatial combination areas of Chinese urban system with different values of Distance Friction Coefficients b. Concerning Chinese urban system, two major regions (1st grade urban system) and seven sub-regions (2nd grade urban system) have formed, the Southern Major Region centralized with Shanghai and the northern one with Beijing. The Southern Major Region consists of four 2nd grade regional urban systems including the Shanghai System (Shanghai, Jiangsu, Zhejiang, Anhui and Fujian), Wuhan System (Hunan, Hubei and Jiangxi), Guangzhou System (Guangdong, Guangxi and Hainan) and Chongqing System (Chongqing, Sichuan, Yunnan, Guizhou and Tibet). The Northern Major Region consists of three 2nd grade regional urban systems including the Beijing System (Beijing, Tianjin, Hebei, Henan, Shandong, Shanxi, Shaanxi and Mongolia), Shenyang-Harbin System (Liaoning, Jilin and Heilongjiang) and Lanzhou-Urumqi System (Gansu, Ningxia, Qinghai and Xinjiang). Besides, 64 local urban systems (3rd grade) have formed in the Chinese urban system, among which the 3rd Grade Shanghai System covers Wuxi, Ningbo, Hangzhou and Jinhua Systems, and the 3rd Grade Beijing System includes Beijing System and Datong System.

客源市场预测是旅游科学决策的关键,由于缺乏历史数据,对尚未开业的旅游景区客源市场预测成为研究的难点.上海迪士尼乐园投资巨大,影响深远,广受关注,开园在即,迫切需要学界从理论层面对其客源市场展开研究.遵循"要素选取--系统分析--模型构建--市场预测--模型验证"的研究思路,对已有旅游引力模型进行修正:① 引入出游意愿,用百度指数进行测度,以明确旅游目的地偏好;② 引入出游率,用旅游人口取代总人口,以界定客源规模基数.而且修正模型中解释变量之间没有显著相关.基于2009-2013年香港迪士尼乐园内地客源市场数据,中国各省市区以及重点城市社会经济数据,运用修正引力模型,预测上海迪士尼乐园国内客源市场空间结构.研究表明:① 迪士尼乐园的强大品牌效应,在一定程度上弱化了空间阻尼的影响,进而导致内地游客对迪士尼乐园的旅游需求整体上缺乏弹性;② 上海迪士尼乐园国内客源市场在空间分布格局上,存在明显的近域指向,东部指向和大中城市指向特征;③ 经过交互验证,用修正引力模型推算的理论值与实际值的吻合程度更高,预测效果更好.

<p>The market forecast plays a key role in tourism decision-making. However, due to lack of historical data, it remains an unsolved problem for the tourist market forecast of destinations which are not yet open to the public. Shanghai Disneyland has attracted wide attention for its huge investment and profound influence. With its forthcoming development, it is imperative to carry out a theoretical research in this field.</p> <p>The existing tourism gravity model is presented with three main explanatory variables: attractiveness of tourist destination, emissiveness of tourist origin, and spatial damping between the destination and origin. This paper makes modifications on the model as follows: (1) It introduces tourist rate which aims at replacing total population with tourist population to measure the tourist scale and accurately determine its base quota. (2) It introduces the element of tourist willingness, which is measured and estimated by using Baidu Index to clarify the preference for tourism destination. Thus, the scope of applying gravity model is not solely confined within the large and medium-scale tourist destinations (namely, cities, provinces and countries). The modified gravity model can also be applicable to the small-scale tourist destinations (namely, scenic spots). Likewise, it will avoid the interference of intervening opportunity. Furthermore, there is no significant correlation between any two variables in the modified model.</p> <p>Based on statistics of mainland tourists to Hong Kong Disneyland and socioeconomic statistics of China's provinces and municipalities from 2009 to 2013, this paper forecasts the spatial structure of domestic tourist market of Shanghai Disneyland through the modified gravity model. The research shows that: (1) Disneyland's strong brand effect drives tourists to overcome obstacles, so the impact brought about by spatial damping is reduced to some degree; as a whole, tourism demand of mainland tourists for Disneyland is inelastic. (2) The spatial distribution of domestic tourist market of Shanghai Disneyland remains strongly concentrated in neighboring regions accoding to the distance-decay theory. It is also mainly concentrated in eastern region and major cities driven by level of socioeconomic development. Based on the forecast of this paper, the Yangtze River Delta region is expected to account for 71.45% of the domestic market share, and China's eastern region occupies 82.40%, 39 major cities 41.44%. (3) Through verification, the coincident degree between the theoretical value deduced from the modified tourism gravity model and the actual value is better than that between the theoretical value deduced from the existing model and the actual value. Accordingly, the newly modified tourism gravity model proves to be more effective than the existing model.</p>

客源地与目的地之间的旅游空间相互作用是必然而持久地,这深刻地影响着旅游者行为和旅游业进步,需要发展适宜的引力模型对其进行刻画和量度.长期以来,旅游引力模型在对空间阻尼的处理上类比牛顿型的幂函数衰减模式,难以克服一些固有缺陷,因此重新回到威尔逊型的指数函数衰减模式就成为一种可能选择.本文即是基于威尔逊形式,将目的地“吸引力”、客源地“出游力”和两地间的“空间阻尼”作为3 个基本解释变量类型,构建了一个基础的旅游引力模型,并重点在参数估计和模型应用上进行了初步探索.由于难以从整体上对模型的两个核心参数进行回归估计,本文首先从局部对收入弹性系数α进行回归估计,然后通过“口粒子模式法”和“出游量积分法”的引入和交互验证来确定空间阻尼系数β的合理取值.研究表明,就全国平均而言,在21 世纪初收入弹性系数可取0.64,而空间阻尼系数可取0.00322,且在空间阻尼影响下,中国大陆居民国内旅游的理论出游半径均值约为300 km.最后,本文从目的地供给视角计算了中国大陆分省区的旅游吸引力,从客源地需求视角计算了成都市到访游客的分省预期市场份额,这两个案例研究在取得一些有益发现的同时也表明威尔逊型旅游引力模型具有积极的应用潜力,值得进一步探索.

Spatial interaction between tourist origin and destination is a key factor affecting tourist behavior and tourism industry. Usually, such a spatial interaction was described by gravity models. However, tourism gravity models used to adopt power deterrence function to describe the spatial friction effect, which is an analogy with Newton's gravity model, are hard to overcome some inherent defects. Therefore, Wilson's model with exponential deterrence function becomes a possible alternative. Based on Wilson's model, a basic form of tourism gravity model is presented with three main explanatory variables: attractiveness of tourist destination, emissiveness of tourist origin, and spatial damping between the destination and origin.<br> Two coefficients, α (income elasticity) and β (spatial damping) in this model are also need to be evaluated. We used the traditional regression method to estimate the value of α. As to β, two new methods, "population particle pattern method" and "integral method on tourist amount" are used to estimate it. The results show that: 1) α = 0.64 and β = 0.00322 are at the national average level in the 2000s; 2) α becomes larger as the field-pixel becomes smaller. For provincial, municipal, county, and township levels, the values of β are 0.00044, 0.0014, 0.0044 and 0.014, respectively; 3) affected by spatial damping, the average travel radius of domestic residents is about 300 km.<br> By the application of this model, attractiveness of each province of China and provincial tourist market shares of Chengdu city are calculated. The results show that: 1) from 2004 to 2008, the average tourism attraction of Sichuan, Liaoning and Yunnan rank the top three, while Ningxia, Qinghai and Inner Mongolia rank the bottom three; 2) from 1999 to 2008, in terms of tourism attraction at provincial level, Xizang has the biggest increase in the ranking, while Shanghai has the greatest decline in the ranking. 3) The results of theoretical calculation on Chengdu city are in accordance with empirical experiences and sampling data, which demonstrates the applicable potential of the tourism gravity model proposed in this paper.

目的地竞争模型是空间相互作用领域的重要进展之一,但其有效性尚未得到一致认可,且缺乏基于中国的实证依据。基于中国2010年城市间铁路客流数据,采用目的地竞争模型进行实证分析,并与传统空间相互作用模型相比较,以检验目的地竞争模型在实际应用中的有效性。结果表明：① 空间结构对中国城市间铁路客流存在显著影响,目的地之间存在较强的竞争效应;② 目的地竞争模型的引入显著地减弱了距离衰减参数的空间自相关程度,较大程度上改善了传统空间相互作用模型的距离衰减参数标定偏误问题;③ 既有研究中在区域尺度下对传统空间相互作用模型（即重力模型）参数的标定及实证分析可能会存在偏误,目的地竞争模型这一改进模型具备应用价值。

Spatial interaction model is an important research field. Existing studies indicate spatial structure of destinations has a significant impact on spatial flow. Thus, traditional spatial interaction models suffer model misspecification problem because the absence of spatial structure variable. Among the modified models introduced to solve the misspecification problem, the competing destinations model is the most widely-used one.The competing destinations modelassumes that the travelers’ destinations selecting process adopts a hierarchical information processing strategy.Based on this strategy, the spatial decision process is divided into two stages. In the first stage, travelers select a destinations cluster containing a set of destinations; in the second stage, travelers select an individual destination from the cluster selected in the first stage. The competing destinations model has been empirically applied in numerous studies in foreign countries.However, the empirical conclusions with respect to the validity of the competing destinations model are still far from agreement. Moreover, none empirical study of this model has been conductedin China. This study applies the competing destinations model based on intercity railway passenger data in 2010 in China, and test its validity by comparing it with traditional spatial interaction models. The estimations of the competing destinations model as well as the traditional spatial interaction model are conducted by the maximum likelihood method, which is calculated by a new method distinguishing from existing studies, i.e. the Particle Swarm Optimization (PSO) algorithm. The conclusions can be drawn as follows. 1) Spatial structure has a significant impact on intercity railway passenger flow of China, and there exists a significant competing effect among destinations both in the system-wide estimation results and in the origin-specific estimation results. The system-wide distance-decay parameter estimated in the competing destinations model (-1.165) is more negative than in the traditional spatial interaction model (-1.108). In the other hand, 124out of a total number of 177 (the ratio is 70% ) origin-specific distance-decay parameter estimationsare more negative in the competing destinations model than in the traditional spatial interaction model, while 140 out of 177 (79%) origin-specific destinations accessibility indicator estimations are negative in the competing destinations model. These characteristics have not ever been reported in Chinese context in existing studies. 2) The competing destinations model reduces the spatial autocorrelation among distance-decay parameters, thus significantly corrects the misspecification problem of traditional spatial interaction models. These results illustrate that the competing destinations model performs significantly better than the traditional spatial interaction model, and thus the improvements by the competing destinations model are empirically valid in Chinese context. 3) The parameters estimation and empirical analysis of traditional spatial interaction models (i.e. gravity model) in existing literature may be biased, while the competing destinations model is an efficient improvement and can play an important part in empirical analysis.

构建城市网络化与消除行政边界所导致的阻隔效应是城市群一体化过程中所面临的双重问题。网络视角下,行政边界呈现出何种程度的空间效应是揭示区域演化机制的科学问题。基于人口流动要素构建长三角城市群网络,定量识别城市网络中的省际边界效应问题。研究指出：① 网络密度、平均中心性和平均控制力三项指标一致表明基于人口流动的长三角城市群网络呈现出强联结状态;② 经济规模、政治资源、产业结构、历史基础、就业机会和劳动收益是人口流动网络形成的重要影响因素,而省际边界、跨界频数、时间距离是人口流动网络流向的关键引导因素;③ 基于二次指派程序（QAP）的人口流动影响因素和网络拓扑结构分析共同表明,长三角城市群网络中存在显著的边界效应,且人口流动网络中边界效应表现出明显差异。省际边界在长三角城市群人口流动网络中发挥了融合（沪苏边界上海-苏州段）、过滤（沪浙边界上海-嘉兴段等）、接触（浙皖边界湖州-宣城段等）、阻隔（苏皖边界常州-宣城段等）四种不同的空间效应。论文对于深入理解城市群一体化驱动机制具有理论和现实意义。

Construction of urban network and the barrier effect (caused by the elimination of administrative borders) are problems in the integration of urban agglomeration. From the perspective of network, the extent of spatial effect indicated by administrative borders reveals the mechanism of regional evolution. This study developed the Yangtze River Delta urban agglomeration network based on the population flow elements, and also quantitatively identified the problems of inter-provincial boundary effects in the urban network. The results showed that: (1) Network density, average centrality, and average control force indicate that the Yangtze River Delta urban agglomeration network shows a strong connection, based on the population flow. (2) Economic situation, political resources, industrial structure, historical basis, employment opportunities, and labor gains are important influencing factors in the formation of a population flow network, while inter-provincial border, cross-border frequency, and temporal distance are key guiding factors of the flow direction of the population flow networks. This conclusion also just proves that the traditional elements group is not able to fully explain the formation and development mechanism of urban network under the background of elements flow efficiently. (3) The analysis of influencing factors of population flow, based on the quadratic assignment procedure (QAP) and network topological structure, showed a significant boundary effect in the Yangtze River Delta urban agglomeration network. Furthermore, the boundary effect of population flow network showed obvious differences. The provincial boundary considers four spatial effects in the Yangtze River Delta urban agglomeration network, including: integration (the Shanghai-Suzhou section on the border between Shanghai and Jiangsu), filtration (e.g. the Shanghai-Jiaxing section on the border between Shanghai and Zhejiang), contact (e.g. the Huzhou-Xuancheng section on the border between Zhejiang and Anhui), and barrier (e.g. the Changzhou-Xuancheng section on the border between Jiangsu and Anhui). This study has theoretical and practical significance for further understanding of the driving mechanisms of urban agglomeration integration.

交通流是城市经济联系最显性的表现,为更清晰地识别城市网络结构提供了新视角。本研究采用高速公路联网收费数据作为交通流&#x0201c;关系数据&#x0201d;,综合图论、社会网络分析法（SNA）和GIS空间分析,以广东省为研究区域探索高速公路交通流网络空间结构特征,并利用QAP回归法探究影响网络形成的因素。结果表明：① 广东省高速公路已形成典型的网络式空间结构,具有&#x0201c;小世界&#x0201d;网络效应。其节点度值服从幂律无标度分布,交通出入量基本平衡,节点重要性自区域中心向外围递减;网络有向边权值等级差异显著,优势交通流具有&#x0201c;向心指向&#x0201d;特征和距离依赖性。② GDP、常住人口、社会消费品零售总额、县区间行驶距离等已成为影响广东省高速公路交通流网络形成的重要因素。其中人口因素是影响城市间交通流动的主要因素,网络形成受制于距离因素,消费因素逐渐成为刺激城市间交通流动的强大动力。③ 广东省高速公路交通流动模式分为邻近指向型、中心指向型和等级指向型三种,以邻近指向型为典型模式。研究结果对于揭示交通流动规律,丰富城市网络理论,推动道路合理规划和区域协调发展具有重要理论和实践意义。

Traffic flow is the most powerful indicator to map urban economic connections, which provides a new perspective to identify urban network spatial structure clearly. This study uses Guangdong's data to explore the spatial structure of expressway network based on theory of synthesized graph, social network analysis (SNA) and GIS spatial analysis. QAP regression method is used to explore the factors which influence the network characteristics. The results show that: (1) The Guangdong's expressway forms a typical network spatial structure with “small world” network effect. The out-degree and in-degree values of network nodes follow the power law scale-free distribution, traffic volume is basically balanced, and the importance of nodes decreases from regional center to periphery. There are significant differences in the weights of directed edges, and the dominant traffic flows have the characteristics of “centripetal direction” and “distance dependence”. (2) The results of QAP regression show that the character of Guangdong's expressway traffic flow network is significantly affected by GDP, resident population, total retail sales of consumer goods, and the distance. Population is the most powerful factor which affects inter-city traffic flows, the formation of network is restricted by distance, and the consumption factor is gradually becoming a powerful driving force to stimulate the inter-city traffic flows. (3) The traffic flow patterns of expressway in Guangdong are divided into three types, namely, proximity-oriented, central-oriented and hierarchical-oriented. The proximity-oriented traffic flow is typical, which is widely distributed in the province; the central-oriented traffic flow is concentrated in the Pearl River Delta urban agglomeration; and the hierarchical-oriented traffic flow is mainly distributed between Guangzhou and Qingyuan, and between Shenzhen and Shantou. The research results have important theoretical and practical significance for revealing the laws of traffic flow, and enriching urban network theory. Furthermore, it provides theoretical support for scientific road planning and coordinated regional development.

随着经济全球化和区域一体化的快速发展,城市间的交流日益密切,要素的流动性和互赖性促进了城市网络的形成,并成为一种新的区域组织模式和空间结构。本文基于2018年腾讯人口迁徙数据,构建了372×372关系数据矩阵,并从不同交通方式出发,系统刻画了中国城市网络格局。研究表明：① 网络关联度高的城市主要集中在胡焕庸线以东区域,尤其是长三角、珠三角、京津冀、成渝等城市群地区的聚集程度最高,成为中国城市网络格局的主控力量;而西北半壁的城市处于被支配地位。② 依据迁徙量,将中国城市网络划分为国家级、大区级、区域级、地方级和基座级网络。城市网络结构随交通方式而变化,当网络等级下移时,节点城市趋于增多,网络密度加大,但网络范围趋于缩小,网络等级与迁徙路径存在着密切的关联性,国家级网络与空运相关,区域级网络与铁路运输相呼应,地方级网络与汽车运输相关。③ 不同运输方式的经济时空距离决定了网络结构,是引致网络随路径不同而产生分异的基本因素。

With the rapid development of economic globalization and regional integration, the connection between cities is increasingly close. The mobility and interdependence of elements have promoted the formation of city network and become a new regional organization model and spatial structure. Based on Tencent's migration data, this paper constructs a 372×372 relational data matrix, and systematically depicts the city network pattern in China from different modes of transportation. The results show that cities with high network correlation degree are mainly concentrated in the area east of the Hu Huanyong Line, especially in the Yangtze River Delta, Pearl River Delta, Beijing-Tianjin-Hebei region, Chengdu-Chongqing region and other urban agglomeration areas, which have the highest concentration and become the main control power of China's city network pattern; and the cities in the northwest half are at a disadvantage status. According to the amount of migration, the Chinese city network is divided into national, large regional, regional, local and pedestal networks. The city network structure changes with the mode of transportation. When the network level moves down, the number of node cities tends to increase and the network density increases, but the network range tends to shrink. There is a close correlation between the network level and the migration path. National-level network related to air transportation, regional network related to railway transportation, and local network are relevant to automobile transportation. The economic space-time distance of different transportation modes determines the network structure, which is the basic factor that causes the network to differentiate with different paths.

本文通过构建人口省际流动的关系矩阵,透视2015年中国春节期间人口省际流动的动态过程、网络特征,并对各省人口流入、流出的规模差异进行测度。研究结论如下：①春节前后中国各省每日人口净流入人次存在显著的规律性特征,2月13-17日和2月25日-3月1日为各省节前返乡流和节后返工流期间人口净流入、流出峰值时段,可作为基于春运研究中国人口省际流动的重要依据;②识别了14个人口净流入省、17个人口净流出省,净流入位序前六省和后八省分别吸纳和提供了全国9成人口的集聚与扩散。人口主要流入省的沿海绵延和主要流出省的中部“人口塌陷”共同构成了双纵格局;③各省流出首位流的指向特征显著,南方人口流出省的流出指向存在共性,京津、长三角、珠三角地区的人口集聚路径存在差异。④移动数据为发掘人口流动过程中蕴含的丰富信息提供了平台,基于节后、节前净流入值差值的方法可以有效识别中国人口流动的规模差异、属性特征。通过与以往研究对比,证实了移动数据与基于人口普查数据研究的诸多结论相似性。

China is one of the most active areas of the world's population mobility. The social structure of China during the socioeconomic transition period, the country's development phase, and its unique cultural background together form the Spring Festival travel rush, a social behavioral phenomenon with a significant regularity and a high degree of uniformity and unity. By constructing the 2015 Spring Festival interprovincial population flow relation matrix, we examined the dynamics of population flow and its spatial characteristics. The results are as follows: (1) 13-17 February and 25 February-1 March were the peak population flow periods before and after the holiday season. Inflow and outflow of population between provinces during these time periods can be indicative of interprovincial migration of floating population in China. We identified 14 net population inflow provinces and 17 net population outflow provinces. The top six and bottom eight population inflow provinces in the eastern and central regions form the double vertical pattern of immigration and emigration of floating population in China. (2) Provincial population outflow primary directions are clear and flow from the central to the eastern coastal areas is the main direction and path of migration of the floating population. Guangdong and Beijing are the primary migration destinations of the floating population in southern and northern china. These two province/municipality monopolized 2/3 of the interprovincial population flow of the country. (3) The source areas of the floating population in the Beijing and Tianjin area, the Yangtze River Delta, and the Pearl River Delta—the three major population agglomeration areas—are significantly different. Those in the Beijing and Tianjin area and the Pearl River Delta are mainly directly from the floating population emigration provinces, but the Yangtze River Delta has formed a more advanced network structure. (4) Mobility-based study on the temporal and spatial characteristics of China's population flow contains a wealth of information on floating population migration, and the Spring Festival travel rush provides an opportunity for such study. By comparing the result with previous research results, similarity between the new data and many conclusions based on the census data is clear.

重力模型是重要的空间相互作用模型,已经在相关研究和实践中得到了广泛应用。在重力模型应用中,结果对参数的敏感性较高,不同的参数设置就可能导致结果体系的巨大差异,故精确的参数是重力模型应用的基础。许多研究关注了重力模型参数的提取,但限于数据等因素,参数在空间上的差异性尚未得到充分的重视。本文采用2010年城市间铁路客运量作为城市间相互作用强度的反映,以城市市辖区居民总可支配收入作为城市的&#x0201c;质量&#x0201d;,刻画城市本身的吸引力,进而通过回归分析考察重力模型的参数结果;并结合地理信息系统的空间插值工具,分析重力模型中各个参数的空间差异并得到可视化结果。相比于既有研究,本文在变量选取中增加了城市间列车交流频次这一变量,与平均运行时间结合共同反映城市之间的&#x0201c;距离&#x0201d;,使模型拟合结果得到较大改进。结果表明,重力模型参数取值在空间上存在显著差异,对不同区域使用统一的重力模型参数可能使结果出现较大偏差,因此将重力模型应用到具体实践中时,不能忽略参数的空间异质性。

As a useful spatial interaction modeling method, gravity models have been widely applied in geographic research. The results of gravity models are sensitive to model parameters, which may lead to significant variation in modeling results when different sets of parameters are used. Therefore accurate parameters are essential to the application of gravity models. Although there have been various studies on the calibration of gravity model parameters, spatial heterogeneity of model parameters has not been carefully considered due to the lack of necessary data for such analysis. In this study we took the volume of China's railway passengers between city pairs in 2010 as the indicator of the magnitude of spatial interaction between cities. Nodal attraction of cities was depicted by the total disposable income of residents in the city areas, which represents the attraction of a city. Compared to the existing literature, this study modified the basic form of gravity model and introduced frequency of trains as a variable into the model, representing the spatial separation between city pairs along with travel time. This adjustment produced improved goodness-of-fit result. Three parameters in the modified gravity model were tested: nodal attraction, α; travel time, β; and frequency, γ. Spatial heterogeneity of each parameter was analyzed by means of regression analysis and GIS. The results show that β has the largest range, followed by α and γ. The research indicates that remarkable spatial heterogeneity exists and cannot be overlooked. Using a single set of parameters of gravity model for different regions may not be appropriate.

<p>Distance is always an important parameter of spatial interaction models in geography and influences the range and depth of many social and economic activities. There is a close and important linkage between traffic flow and distance. So the regularity of distance decay of traffic flow becomes the important research field of transportation geography and economic geography. Based on the review and comment on research progress of distance decay regularity of traffic flow in foreign countries and China, we narrated the studying data and city specimens applied in this paper, and introduced the primary modals of distance decay function, including exponent model, Pareto model, logarithm model, square exponent model and extraction exponent model. Then we simulated the five distance decay model applying inter-urban traffic flow in China, including road passengers flow, railway passengers flow and railway cargo flow, and aviation passengers flow. We discussed the several parameters of their simulating results including distance decay coefficient b, correlation coefficient R and test parameters (namely F and T), especially provided a comparative analysis. According to the analytical result and simulating parameter, we chose the optimal function model of distance decay of traffic flow, namely, Pareto model, and wrote the specific function formulations of the four traffic flow modals.&nbsp; So we further analyzed the parameter difference of four distance decay models and their geographical implication, especially the relation between distance decay coefficient b and the application distance of four traffic flow modals, and their changes about distance decay parameter of temporal inter-urban railway passengers and cargoes flow and their spatial explanation. To provide a further explanation, it is necessary to do so through analyzing the traffic flows and related social-economic index. So we protracted the accumulative ratio curves and distribution ratio curves of four traffic flows, and demonstrated the distance decay model of traffic flow, their parameter and decay regularity which provided an experiential discussion.</p>

步行不仅是一种原始、便捷的交通方式,同时也是体力活动的重要组成部分,对于提升公共健康、改善交通拥堵和减轻污染排放等均有重要的积极意义。然而,包括青少年在内的城市居民步行比例持续下降,体力活动水平日益降低。青少年正处于身心发育的关键时期,体力活动的缺乏将导致肥胖等慢性非传染病,为其将来发展埋下巨大的健康隐患。如何通过对建成环境进行干预,提高青少年步行通学比例,从而提高其体力活动水平,引起了不少学者的关注,取得了较为丰硕的研究成果。然而,既有研究存在以下不足：第一,大部分已有研究以西方城市为案例,很少研究关注中国城市;第二,绝大部分既有研究基于线性或广义线性的假设考察建成环境对步行通学的影响,很少研究关注两者之间的非线性关系。鉴于此,论文以厦门岛为案例,基于极限梯度提升模型,考察青少年家和学校建成环境对其步行通学的影响。研究发现：① 通学距离是影响青少年步行通学最重要的因素,其相对贡献接近4成(39.99%);② 建成环境(以5Ds模型表征)作用显著,家、校建成环境相对贡献合计达36.28%,超过社会经济属性(23.73%),离市中心的距离和道路交叉口密度等变量具有重要作用;③ 全部建成环境变量和主要社会经济属性变量均与青少年步行上学存在非线性关系,且存在明显的阈值效应。研究为城市决策者关于提高青少年步行通学倾向提供了丰富的政策启示。

Walking is not only a primitive and convenient transport mode but also an important integrant of physical activity, which is beneficial for the promotion of public health, alleviation of traffic congestion, and mitigation of transportation-induced pollution. In modern China, cities are expanding rapidly, people are enjoying a dramatic improvement in living standards, and the pace of life is accelerating. As a result, urban people, including adolescents, tend to travel in motorized modes increasingly more and walk less. The prevalence of physical inactivity among adolescents has brought about a series of health issues, such as deterioration of physical fitness, obesity, and some non-communicable diseases (for example, diabetes and hypertension). Travel to school is among the most important routine travels for adolescents. Promoting adolescents' propensity of walking to school can effectively help them integrate physical activity into daily life and thus enhance their overall physical activity level. Hence, scholars from diverse disciplines (for example, geography, urban planning, and public health) have been drawn to examine the relationships between the built environment and walking to school. However, the current research is insufficient in the following two aspects. First, the existing research is mainly based on the Western context, whereas few studies have been conducted in China. Second, the majority of existing studies assumed a linear or generalized linear (for example, log-linear) relationship between the built environment and walking to school, and no studies, to the best of our knowledge, have examined the non-linear relationships between them. Therefore, this study, taking Xiamen, China as the case and employing its large-scale travel behavior survey dataset in 2015, explored the non-linear effects of the built environment on adolescents' propensity of walking to school. We applied a state-of-the-art machine learning method, namely extreme gradient boosting method (XGBoost), to fit the model, and interpreted the model with relative importance and partial dependence plots. The results show that: 1) Distance from home to school is the most important factor influencing walking to school, with the relative importance of 39.99%. 2) The built environment, which is characterized by the 5Ds (density, diversity, design, destination accessibility, and distance to transit) model, is an important contributor, and relative contributions of the built environment variables at home and school collectively contributed 36.28% of the model's explanatory power, only second to distance to school, much higher than that of sociodemographic variables (23.73%). Distance to city center and population density around both home and school contribute a great deal. 3) All the built environment variables at both ends of school trips and the key sociodemographic variables have non-linear effects on adolescents' propensity of walking to school, and there exist obvious threshold effects. This study can inform decision makers with nuanced policy insights for promoting adolescents' behavior of walking to school.

城中村作为带着农民特色和阶级混杂重塑的空间,为弱势劳动群体融入城市化体系、实现反资本而不反市场的城镇化进程提供了保障。探讨建成环境对居民通勤方式选择的影响机制,开展城中村建成环境改造,对于解决弱势群体通勤和就业困难意义重大。论文基于武汉市2018年LBS(Location Based Services)数据,以公共交通通勤方式选择作为二元分类因变量,采用梯度提升决策树(gradient boosting decision tree,GBDT)分析建成环境与城中村居民通勤方式选择的非线性关系与阈值效应,结果显示：普遍长距离通勤的城中村居民对公共交通系统依赖性强,建成环境的非线性影响和阈值效应明显,弥补了传统线性模型在可解释性上的不足;其中公交站可达性预测贡献度最大(19.89%),阈值分别在300 m和1000 m左右,在城中村改造时应尽量满足直线距离300 m公交站点可达,保证1000 m内存在公交线路,其余变量也有相似的可解释性。研究得到的建成环境影响的优先层次和最有效的空间属性范围将为提升城中村居民就业和通勤公平,实现区域包容、协调的城镇化提供参考。

As a space with farmers' characteristics and reshaped by mixed classes, urban villages provide a guarantee for the vulnerable groups to integrate into the urbanization system and realize the anti-capital but not anti-market urbanization process. Exploring the impact mechanism of the built environment on residents' commuting mode choices and carrying out inclusive renovation of the built environment of urban villages will be of great significance for addressing the commuting and employment difficulties of vulnerable groups. Based on the location based services (LBS) data of Wuhan City in 2018 and taking the choice of commuting mode as binary dependent variable, this study used the gradient boosting decision tree to explore the nonlinear relationship between the built environment and commuting mode choices of urban village residents and threshold effect. The results show that the residents of urban villages who commute for a long distance generally depend on the public transport system, and the nonlinear influence and threshold effect of the built environment are obvious, which make up for the insufficient interpretability of the traditional linear model. The contribution of bus station accessibility is the largest (19.89%) in prediction, and the thresholds are about 300 m and 1000 m respectively. In urban village reconstruction, bus station within a Euclidean distance of 300 m should be satisfied as far as possible, there must be bus lines within 1000 m, and other variables have similar interpretability. The impact priority level and the most effective spatial attribute range of the built environment obtained from this study will provide a reference for improving the employment and commuting equity of urban village residents and for realizing regional inclusive and coordinated urbanization.