Investigate Jobs–Housing Spatial Relationship with Individual-Based Mobility Big Data of Public Housing Tenants: A Case Study in Chongqing, China

Abstract

Dwelling and working are two of the most fundamental urban functions. The jobs–housing relationship is pertinent to a city’s spatial and social structure. By using macro-level statistical data and micro-level mobile phone data, this paper innovatively adapts a way to explore the jobs–housing characteristics of public housing tenants in Chongqing, China. It finds that both external and internal factors matter. This study might provide some useful implications for further policy-making on public housing planning, and the research method could be used to examine urban spatial relationships further.

1. Introduction

After rapid development for decades, urbanization in China has been entering the second-half stage, which is human-oriented. It aims to solve the conflict between the gradually increasing demand for a better life and the insufficient and imbalanced social development. In the urban housing field, some low-income residents feel it is harder to afford rapidly rising housing prices. Under this circumstance, local governments in China have begun to develop public rental housing, aiming to improve the living condition of people with housing difficulties. However, with the completion and use of public rental housing projects in China, people find the problems of remote location, inconvenient transportation, and the consequent residential isolation and social marginalization of the tenants. The jobs–housing relationship of public rental housing will increasingly impact urban land use, transportation, commuting, job self-sufficiency, etc. There are relatively complete studies on public rental housing and jobs–housing spatial relationships. However, few have examined the jobs–public rental housing relationship. From this perspective, this paper focuses on the jobs–housing characteristics of people living in public rental housing.

Research on dwelling and working spatial relationships can be traced back to Howard’s “Garden City” at the end of the 19th century. Due to rapid urbanization, London suffered from severe urban problems such as overcrowding and chaotic traffic. Howard believed it to be an effective way to solve these problems by establishing new cities with well-developed service facilities nearby London. He advocated that the dwelling and working places should be located close to each other to “working within walking distance of the residence”. Later, Mumford further clarified the connotation of “balance”. He believed the balance of various functions within the city could be achieved by controlling population, adjusting residential density, and planning urban areas [1]. In the 1960s, under the circumstance of racial discrimination in the housing market and the gradual suburbanization of jobs, the theory of “spatial dislocation” came into existence. The theory pointed out that the “jobs–housing mismatch” between the number of jobs and job seekers was an important reason why employers in suburban areas were unwilling to hire employees (mostly African Americans) living in inner cities [2]. This social phenomenon reflected that the rapid spatial restructuring might affect disadvantaged groups’ living and employment opportunities. Since then, this creative hypothesis has been extended to various fields of urban studies. Researchers in related fields such as urban planning and sociology have carried out many empirical studies on the spatial characteristics of women, ethnic minorities, and low-income and middle-income groups. As an ideal model of the dwelling and working spatial relationship, “jobs–housing balance” is a planning concept gradually formed by western planners when struggling against “urban diseases” [3]. It generally believes that “jobs–housing balance” is beneficial to reducing commuting time, alleviating traffic congestion, and promoting the sustainable development of cities [4,5].

With the economic growth, institutional transformation, and spatial restructuring in Chinese cities, Chinese scholars began to study the spatial relationship between dwelling and working in the mid-1980s. Concerning the evolution of urban spatial structure, scholars mainly obtain data from the macro- and meso-levels for the studies [6,7]. At the macro level, the scholars take the whole city as the research object. These studies use the population and economic census data and take the streets, districts (counties), and zip code communities as spatial units. For example, Cheng and Tang discovered that the permanent population and employment positions in the central urban area of Shanghai showed the evolutionary characteristics of separation from each other [8]. At the meso-level, the studies use questionnaire data such as individual characteristics and commuting methods of residents. Wang et al.’s research explored the transformation of the jobs–housing relation from spatial dependence to spatial mismatch in Beijing [9]. Hu and Zhu used questionnaire data to construct a series of indicators and found that Wuhan was undergoing a centripetal urbanization development [10]. In recent years, with the rapid development of information technology, big data has been widely applied in related research fields. Using data such as government data [11], bus card swiping data [12], cell phone signal data [7,13,14,15,16], Baidu heat map data [17], etc., more and more studies on the jobs–housing relationship have been carried out in cities such as Beijing, Shanghai, Wuhan, Chongqing, and so on. In addition, Zhang and other scholars tried to combine big data with small data. Using economic census data and mobile phone data, they found that Beijing has an apparent imbalance between dwelling and working places [18].

Based on descriptive characteristics of the jobs–housing spatial relationship, there are also a lot of studies focusing on the underlying factors leading to jobs–housing patterns. Studies have revealed that the individual characteristics of residents themselves are an essential factor affecting the jobs–housing relationship. Different groups of people have noticeable different jobs–housing characteristics [19]. Other factors such as the availability of children’s picking-up services, commuting costs, and household registration systems also significantly impact the degree of imbalance between the dwelling and working [20]. Scholars such as Zong took the Yuzhong District of Chongqing as an example and found that commuting distance is related to urban functional layout, topographical conditions, and traffic conditions [21]. In contrast, others believe that traffic conditions are not the main factor causing traffic jams in Chongqing’s main urban area at rush hours because Chongqing’s overall jobs–housing pattern is good enough to avoid pendulum commuting traffic [6].

Compared with the above research on the jobs–housing spatial relationship in the whole city, the subdivided groups of public housing tenants who usually earn a low income have attracted less academic attention. Few studies have shown significant differences in jobs–housing spatial relationships and decision-making patterns [22]. Moreover, the underlying factors also vary significantly [23].

In terms of research content, existing studies mainly focus on static descriptions of jobs–housing spatial relationships and the underlying factors and focus less on the patterns of the subdivided groups. In terms of research methods, scholars mostly use self-sufficiency indicators and jobs–housing ratios. Based on them, some indicators have been developed, such as the spatial dislocation index, employment accessibility, the employment and residence dispersion rate, and others [13,16,20,24]. In terms of data sources, most studies use macro-level population and economic census data and meso-level questionnaire survey data. In the era of big data, the data from location-based services provide a new data source to explore the relationship between dwelling and working places.

This paper takes Chongqing, China, as the study area. With macro-level statistical data and micro-level mobile phone data, it explores the jobs–housing characteristics of public housing tenants. It examines whether there is a spatial mismatch and the corresponding social margination. To achieve this purpose, it proposed the following hypothesis: (1) urban spatial structure is an external factor affecting the jobs–housing relationship; (2) the tenants’ internal socio-economic attributes are pertinent to the spatial relationship; (3) nearby employment opportunity also have a direct impact on the jobs–housing relationship.

2. Methodology

2.1. Study Area

At the end of 2019, Chongqing had a permanent population of 31.243 million, of which the urban population was 20.899 million. Affected by topographical factors, most residential areas in Chongqing are distributed in clusters surrounded by mountains and rivers. “Chongqing Urban and Rural Master Plan (2007–2020)” declares that Chongqing aims to develop the spatial form of “one city, five areas, and multiple centers”.

As one of the first cities in the country to pilot public rental housing policies, Chongqing promulgated the “Interim Management Measures for the Public Rental Housing in Chongqing” in 2010 to achieve the goal of “homeownership” for low-and middle-income residents. Public information shows that at the end of 2019, Chongqing has allocated 540,000 public rental housing units to meet the needs of more than 1.4 million people with residential difficulties. Among them, 20 municipal investing projects have all been occupied, with a relatively high occupancy rate (see

Table 1. Public rental housing in Chongqing central city.

Project	Administrative District	Area (million m2)	Residential Units (sets)	Occupied Residential Units (sets)	Occupancy Rate (%)
Peninsula Yijing	Dadukou	135	19,833	19,533	98.49%
Happy Garden	Dadukou	91	/	/	/
Beautiful Sunshine	Shapingba	111	15,255	15,023	98.48%
Xuefu Yueyuan	Shapingba	186	27,051	6952	25.70%
Kangju West City	Shapingba	147	24,330	21,690	89.15%
Min’an Huafu	Jiulongpo	125	19,288	18,688	96.89%
Golden Phoenix Garden	Jiulongpo	67	9616	777	8.08%
Chengxi Home	Jiulongpo	74	10,939	2439	22.30%
Jiulong West Garden	Jiulongpo	107	14,383	2051	14.26%
Chengnan Home	Nan’an	199	32,649	31,817	97.45%
Jiangnan Waterfront	Nan’an	126	/	/	/
Liangjiang Minju	Beibei	107	17,156	16,275	94.86%
Jinyun New Residence	Beibei	82	/	/	/
Beidu Jiayuan	Beibei	89	/	/	/
Airport Paradise	Yubei	185	29,862	27,372	91.66%
Yunzhuan Shanshui	Ba’nan	140	17,466	16,709	95.67%
Longzhou South Garden	Ba’nan	36	/	/	/
Qiaoping Renjia	Ba’nan	87	11,628	6031	51.87%
Minxin Jiayuan	Liangjiang New Area	112	18,238	17,689	96.99%
Kangzhuang Meidi	Liangjiang New Area	135	22,830	21,979	96.27%

Source: Chongqing Municipal Public Rental Housing Administration and public information. The statistical time was on 11 December 2019. The author has compiled it.

). They have formed a relatively mature “Chongqing Model” of public rental housing development, and its policy coverage and construction scale are at the top level in the country.

2.2. Research Method

2.2.1. Data

A fundamental part of researching jobs–housing spatial relationships is recognizing the places of dwelling and working effectively. Current studies mainly use macro-level census data and meso-level questionnaire survey data. The former are less general, and the latter are restricted by time, workforce, and financial resources. In the era of big data, trajectories of residents can be extracted from individual-based mobility data, such as mobile phone data. For each resident, it is possible to figure out frequently visited places. If one place was visited every midnight and stayed for several hours, it is most likely that it is the dwelling place of the resident.

Similarly, if the resident visits a place every workday, it might be his or her workplace. It is noted that, although the above method might not be precise enough, in favor of the merits of big data, it is still possible to locate the dwelling and working places of a majority of regular commuters. By this means, this paper adopts a research method that combines macro-level statistical data and micro-level mobile phone data. The mobile phone data are from a set of databases government released for research only, excluding any sensitive and private information.

This paper uses the number of employees and permanent residents of Chongqing and each administrative region from the Statistical Yearbook as the data source for the number of the employed population and the number of residents in the region. Based on the data, the paper calculates the “jobs–housing ratio” [25] and “jobs–housing deviation ratio” [26] of the nine districts of Chongqing to explore the overall jobs–housing spatial characters in Chongqing. In addition, this paper uses the mobile phone data in nine days of April 2017. Considering Chongqing’s spatial form, the specific construction progress of various public rental housing projects, and the validity of the data, this paper focuses on examining the eight projects (Figure 1). After data processing, this study identified 16,984 samples living in the studied public rental housing.

2.2.2. Data Analysis

Existing studies have shown that the indicators for measuring the spatial distribution of occupation and resident include two perspectives: static characteristic indicators and dynamic correlation indicators. According to the existing characteristics of the data mentioned above, macro-level yearbook statistics were used to study the jobs–housing ratio (JHR) and jobs–housing balance index (JHB), and the micro-level mobile phone data are used to analyze commuting characters.

1.

Static description indicators

JHR is a commonly used indicator to measure the jobs–housing spatial relationship. Cervero defined the JHR as “the ratio of the number of jobs to the number of residents in a given area” and believes that the ratio is between 0.8~1.2 [25]. The formula for calculating JHR is as follows:

JHR_i = J_i/R_i

(1)

In Formula (1), J_i represents the number of jobs in area i, and R_i represents the number of residents in area i.

JHB is a commonly used indicator to measure whether the spatial relationship between the occupation and resident is balanced, that is, the ratio of the proportion of the employed population in a region to the total employed people and the proportion of the resident population in an area to the total resident population. If the ratio is 1, it indicates that the residential and employment functions of the region are well-matched. If the JHB is greater than 1, it indicates that employment functions dominate the area, and less than 1 indicates that residential functions dominate it. The calculation formula of the JHB is as follows:

Z_ij = (Y_ij/Y_i)/(R_ij/R_i)

(2)

Z_ij is the JHB index of area j in the year i. Y_ij represents the number of the employed population in area j in the year i. Y_i represents the total number of employed people in the year i. R_ij represents the residential population in area j in the year i. R_i represents the total number of residents in the entire area in the year i.

2.

Dynamic correlation indicators

“Commuting distance” and “commuting time” are formulated using the bus travel measurement interface provided by Baidu Maps. To calculate the commuting time and distance from the dwelling to the working place, this paper has taken the residents’ required public transportation at 7 a.m. on weekdays as the outbound commuting time and distance no matter what transportation these residents choose. In the same way, it calculates the time and travel distance required by public transportation from working to dwelling places at 5 p.m. on working days as the return commuting time and distance.

This study first identified those living in the eight studied public rental housing projects using mobile phone data. It identified a total of 16,984 signal units (individual). Every single user of the cellular signal was treated as a unit. It calculated the time each signal (individual) stayed at specific grids in minutes. Based on the daily results, it further counted all the non-repeated grids visited by a single user during the residential period and working period for five consecutive working days. Then, it identified a fixed location at 11 a.m. for at least four days as the user’s potential workplace, namely the “commuting center”.

After identifying each individual’s commuting center, this study used Baidu Map’s public travel measurement interface to calculate the commuting character further. Instead of travel distance, we employed time-dependent travel time given by Baidu to measure travel cost. It may reflect the jobs–housing spatial relationship more precisely than existing methods. First, Baidu Map’s bus travel interface could provide a unified measurement standard relative to each individual’s possibly real transportation way. Secondly, it is more in line with the objective reality of public rental housing residents using public transportation as the primary commuting method.

3. Data Analysis

3.1. The Static Description Characteristics

3.1.1. JHR

At the end of 2019, the number of jobs in Chongqing was 16.6816 million, with a permanent population of 31.8784 million. The overall JHR is 0.52, which is lower than the minimum threshold of 0.8 proposed by Cervero, indicating that the overall number of jobs in Chongqing cannot fully meet the needs of the residents. Among the nine central city districts, only Yuzhong District’s JHR is between 0.8 and 1.2, while the other administrative districts are all below 0.8 (Figure 2). It reflects the jobs–housing imbalance in most administrative regions, mostly housing-oriented. Although JHR numerically reflects the matching status of the dwelling and working space, it fails to accurately reflect the employment situation of the residents within the same area. Therefore, further research is needed.

3.1.2. JHB

This study calculated the JHB of each district in Chongqing’s main urban area and divided it into different groups according to the standard proposed by Ying Chenglong [26]: Z_ij < 0.7 indicates that the residential function is dominant, which is called the residential-dominant area. 0.7 < Z_ij < 0.9 is called the residential-based secondary matching area. 0.9 < Z_ij < 1.1 is called the basic matching zone. 1.1 < Z_ij < 1.3 is called the employment-based secondary matching zone. Z_ij > 1.3 indicates that the employment function is dominant and is called the employment-oriented zone (see Figure 3). The results show that the jobs–housing spatial relationship in the nine districts of Chongqing central city is significantly different. Yuzhong’s JHB is 1.79, consistent with the situation indicated by the JHR mentioned above. The JHB of Jiulongpo and Yubei District suggests that this region’s dwelling and working spaces are well-matched. Jiangbei District and Nan’an District are sub-matching areas dominated by residential functions. Dadukou, Beibei, Shajiaping, and Ba’nan District belong to the resident-dominant areas.

3.2. Dynamic Correlation Characteristics

The above static analysis provides an overall understanding of the jobs–housing matching status in Chongqing. It is also the fundamental reality faced by the residents commuting in the eight public rental housing communities studied in this paper. The following is a further analysis of the commuting characteristics of these eight communities.

3.2.1. Commuting Time and Commuting Distance

Overall, the average commuting time and distance in the studied communities are consistent in data (

Table 2. The average commuting time and distance of the public rental housing districts.

Community Name	To Work		Back Home
	Average Commuting Time (min)	Average Commuting Distance (km)	Average Commuting Time (min)	Average Commuting Distance (km)
Chengnan Home	50.32	12.436	50.20	12.545
Yunzhuan Shanshui	53.52	12.827	55.83	13.487
Kangju West City	52.74	9.393	51.13	9.589
Liangjiang Minju	50.45	13.354	55.85	13.611
Peninsula Yijing	63.51	12.614	64.80	12.435
Chengxi Home	50.50	6.839	48.24	6.693
Jiulong West Garden	54.06	8.989	52.20	9.099
Kangzhuang Meidi	47.30	10.950	47.11	10.822

), indicating that the method for determining the working place of the public rental housing residents is reasonable. The analysis results show that the average commuting distance of residents in the eight studied communities is 10.93 km, which is lower than the overall average commuting distance of 12.2 km in Chongqing (2018 data). The average commuting time of the eight studied communities is 52.8 min, slightly lower than the average commuting time of 54 min in Chongqing (2018 data). Among them, the shortest one-way time is about 47 min, and the longest reaches are 63.51 min, indicating that the overall jobs–housing spatial relationship in most studied communities is well-matched.

To better identify the commuting characteristics of the studied communities, this paper further divides the commuting distance into five levels (

Table 3. The overall commuting distance of the studied communities.

Unit: km	<5	5~10	10~15	15~20	>20
Studied communities	21.8%	23.5%	22.5%	16.0%	16.2%
Chongqing	48%	27%		25%

, Figure 4).

In terms of commuting distance, the result shows 0–10 km is the main distance of one-way travel, accounting for more than 50% (except for the Chengnan Home) (Figure 4). The commuting distance of the residents in Yunzhuan Shanshui is relatively long, and as much as 28% of the employed population has a one-way commuting distance of over 20 km. More than half of the people in the eight studied communities have fewer than 15 km of commuting distance. Nevertheless, the proportion of people whose commuting distance is less than 5 km is slightly lower than that of Chongqing.

The commuting time is divided into five levels by 25 min (Figure 5). The results show that each studied community displays different patterns. The highest proportion of travel time in most communities is 25–50 min (seven of the eight communities), followed by 50–75 min. These two types account for more than half of the commuters of each community (some communities account for 80%).

3.2.2. Commuting Center

On the whole, the density of the working population in the eight studied communities is distributed in patches, most of which radiate to the periphery, with the studied communities as the center. The workplaces have significant agglomeration, but each studied community displays different patterns (Figure 6). The working places of most communities are centered on the residential communities, spreading outward in a circle, from the center to the periphery. In contrast, Chengnan Home is an exception. Its residents’ core workplaces are distributed on the east and west sides of the community, while the employment density near the community is low. Its employment density is high in the east, centered on an industrial park, a university, and a scenic park in the west. The difference in the distribution of the working population may be related to factors such as the number of employed population living in the community, the location conditions of the community, and the functional layout of the surrounding areas.

4. Discussion

The above analysis shows that the overall jobs–housing relationship of the public rental housing displays a similar pattern with the whole city. This might suggest that urban spatial structure is an important external factor. The JHRs of the administrative districts where the eight studied communities are located outside the reasonable range of 0.8–1.2 for the jobs–housing balance (

Table 4. The characteristics of the jobs–housing relationship of the studied districts.

Community Name	Administrative District	Characteristics of the Administrative Districts
		JHR	JHB
Chengnan Home	Nan’an	0.42	0.75
Yunzhuan Shanshui	Ba’nan	0.28	0.5
Kangju West City	Shapingba	0.35	0.63
Liangjiang Minju	Beibei	0.31	0.56
Peninsula Yijing	Dadukou	0.34	0.63
Chengxi Home	Jiulongpo	0.56	1.01
Jiulong West Garden	Jiulongpo	0.56	1.01
Kangzhuang Meidi	Yubei	0.55	1.00

). The overall JHR of Chongqing’s central urban area is relatively low, and most of them are residential-biased functions. It somehow suggests that jobs–housing is unbalanced in Chongqing, especially Beipei, Ba’nan, Shapingba, Dadukou, and Nan’an. The number of jobs in these districts may not fully meet the population’s needs living there. Therefore, the studies communities in the above districts have a severe commuting problem. The average commuting time and distance of their residents are significantly higher than the average level of Chongqing in 2018.

On the other hand, Yubei and Jiulongpo District, which are jobs–housing balance areas, can provide sufficient jobs to meet the employment needs of the residents in the region to a greater extent. This helps the residents living in Chengxi Home, Jiulong West Garden, and Kangzhuang Meidi to find jobs nearby. It shows that the overall jobs–housing relationship in the region has an important impact on the commuting status of residents living in public rental housing.

According to the analysis, nearly 80% of residents’ average commuting time in the studied communities is lower than the average commuting time in Chongqing. It might be due to these public rental housing projects’ good location conditions and transportation accessibility (Figure 7). Most of the eight studied communities are close to subway lines, which are close to Line 6. Chongqing Subway Line 6 runs through five administrative districts and connects Chongqing’s three major CBDs. Therefore, near Line 6, these communities could significantly shorten the spatial and temporal distance to the city’s core area.

Consequently, it allows some residents in these communities to find long-distance jobs across different districts. The two surrounding public rental housing communities without subway lines are in Jiulongpo. The driving distance to the city center is more than 30 km. Although the commuting distance of these two communities is not long, the commuting time is generally longer due to insufficient public transportation and restrictions of terrain conditions. Therefore, location conditions and public traffic accessibility are important factors affecting the commuting status of public rental housing residents.

In addition, the public housing tenants also show their special jobs–housing characters. First, the surrounding job centers could place a significant impact. According to the thermodynamic diagrams of the distribution of the jobs, most residents of the studies public rental communities choose the jobs nearby. This might be related to the distribution of the surrounding industrial parks (

Table 5. The industrial districts surrounding the studied communities.

Community Name	Industrial Zone	Rank	Approved Area (ha)	Pillar Industry
Peninsula Yijing	Hengtong International Automobile and Motorcycle City	N/A	N/A	Auto and motorcycle accessories
	Jianqiao Industrial Park	Provincial	1152.88	Machinery, new materials, environmental protection
	Happy City Industrial Park	N/A	N/A	N/A
	Longxin Industrial Park	N/A	N/A	Motorcycle, motorcycle engine
Chengnan Home	Nanshan Scenic Area	N/A	2500	Tourism service industry
	Chongqing Economic and Technological Development Zone	National	N/A	Electronic information, equipment manufacturing
	Dongben Industrial Park	N/A	N/A	Motorcycle
	Changdian Warehousing & Logistics Center	N/A	N/A	Logistics
	Chayuan Logistics Center	N/A	1754.96	Electronic information, equipment manufacturing, pharmaceutical and chemical industry
	Tongsheng Industrial Park	N/A	N/A	N/A
	Midea Group Chongqing Industrial Zone	N/A	53.3	Appliances and machinery midwest manufacturing center
	Jianxin Industrial Park	N/A	N/A	N/A
Kangju West City	Chongqing University Town	Provincial	3300	Advanced talent training center, scientific research and innovation center, international science and technology education exchange center
	Chongqing Xiyong Micro-electronics Industrial Park	Provincial	1708.53	Computer, electronic and communication equipment
	Chongqing Xiyong Comprehensive Bonded Zone	National	832	Computer, electronics
	Jinfeng Electronic Information Industrial Park	Provincial	589	Computer, electronics
	Xinyuan Industrial Park	N/A	26.27	Motorcycle
Liangjiang Minju	Caijia Industrial Park	N/A	4.06	N/A
	Lifan Group	N/A	N/A	Production of engines, general machines, motorcycles, and automobiles
	Yingtian Caijia Gonggu	N/A	17.27	N/A
	Yingtian Optoelectronics Valley	N/A	17.45	N/A
	Tongxing Industrial Park	Provincial	1107.99	Machinery, instrumentation, auto and motorcycle accessories
Yunzhuan Shanshui	China Cloud Education Industrial Park	N/A	270	Combining virtual education with physical education
	Banan District Gymnasium	N/A	0.08	Humanities and entertainment
	Huaxi Culture and Sports Center	N/A	N/A	International top events and high-end performances
Chengxi Home	Xipeng Industrial Park	Provincial	1406.39	Aluminum processing, auto and motorcycle parts, food
	Chongqing Taojia Urban Industrial Park	Provincial	66.67	Mechanical processing and manufacturing, electronic information, high-tech, business logistics
	Jiulong Industrial Park	Provincial	2142.59	Cars, motorcycles, smart equipment
	Chongqing Shuangfu Fruit Wholesale Market	N/A	N/A	Wholesale fruit market
	Runtong Industrial Park	N/A	46.67	General-purpose engines, ATV utility vehicles, small household appliances, and their supporting parts
	Chongqing Taojia Urban Industrial Park	Provincial	66.67	Mechanical processing and manufacturing, electronic information, high-tech, business logistics
Kangzhuang Meidi	Liangjiang New District Internet Industrial Park	N/A	N/A	Internet and software information (More than 200 various Internet companies have settled, with more than 10,000 employees in the park)
	Zhaomushan Science and Technology Innovation City	N/A	400	Software industry center, internet industry park, service trade park, advertising creative industry park
	Chongqing Software Industry Center	National	3000	Internet of Things, industrial internet, big data, 5G, artificial intelligence

Source: author, Catalogue of China Development Zone Audit Announcements (2018 Edition).

). Taking Peninsula Yijing as an example, the spatial distribution of its residents’ commuting center highly coincides with the nearby Jianqiao Industrial Park. Jianqiao Industrial Park is the first municipal-level industrial park in Chongqing that has provided lots of relatively low threshold employment opportunities. This kind of industrial region could significantly impact the job distribution of residents living in public rental housing.

Using the information released by the government, this study analyzed the socio-economic attribute of the public housing tenants in Chongqing from 2015 to 2020 (Figure 8). It shows nearly half of the public housing applicants are single, and more than 60% are female. Their average income is RMB 35,000 per year. A total of 80% of them are below 45 years old. This indicates the so-call “sandwich layer” attribute of this group. They are in the first stage of their career (for example, those new graduates of local universities), having relatively low-income and less burden from the family. Therefore, they are more flexible in choosing the place of dwelling and working. However, subject to limited income, most of their commute depends on public transportation. Thus, the nearby jobs have much attraction for them. According to Wang and Yang’s study, their jobs are mainly automobile making, technology, and the service industry [27]. This is in accord with the industrial structure nearby.

5. Conclusions

This paper combines macro-level statistic data with micro-level mobile phone data to establish a series of indicators to examine the connection between public rental housing communities’ dwelling and working places in Chongqing. It tries to fill the research gap in rental housing, especially public rental housing in China. Additionally, this study applies some innovative methods to use mobile phone data to accurately define the dwelling and working places. The method adopted in this paper is based on previous studies with some methodology innovation according to the data and circumstances of the study areas. The research result might be helpful to grasp a clear map of the jobs–housing relationship of public rental housing. All the three hypotheses proposed are verified. It suggests similarity and difference between the public rental housing communities and the overall pattern. The similarity is decided by the city’s overall structure and transportation accessibility. For the difference, the socio-economic attribute of the public housing tenants and the surrounding jobs center are decisive. In this account, the construction of public residential spaces for tenants in the future should consider the need for a workforce for nearby industrial areas to reduce commuting.

According to the “2018 Housing Leasing White Paper in China”, about 168 million people in China have to meet their living needs via renting housing. In the next five years, the number will increase by 50%. Most tenants are newly employed college graduates and migrant workers whose housing affordability is relatively weak, but employment pressure is relatively high. In this regard, the balance between dwelling and working is crucial. For people living in public rental housing, the residential choice is about income distribution, commuting efficiency, and living amenity. For government management, the location of public rental housing is the compromise of land finance, urban spatial restructuring, social justice, etc. When making related policies, it requires local governments to consider the basic characteristics of public housing tenants particularly and the city’s spatial structure to solve their living difficulties and moreover improve the tenants’ quality of life. The finds of this paper could help the policy-making in public rental housing and urban spatial restructuring. Due to the data limit, this study could not cover all the public rental housing projects in Chongqing and could not compare their jobs–housing pattern with others. Future studies will focus on exploring the underlying driving force of the location choice. The methodology adopted in this study could be applied in the future to study such as the dwelling and schools relationship. It may help in comprehensively depicting the residential choice faced by certain groups of people.