Effects of land and building usage on population, land price and passengers in station areas: A case study in Fukuoka, Japan

==Abstract==

This study uses multiple regression to investigate the effects of land and building use on population, land price, and passengers. Initially, we abstract annual data on land and buildings usage within a radius of 0 m–400 m for railway stations and 400 m–800 m for subway stations in Fukuoka, Japan by using the GIS. We then analyze the relationships between 13 factors of land use and 8 factors of building usage, as well as the related population, land price, and passengers using the quantitative expression method. Using several categories of land use and building usage as explanatory variables, we analyze the degree to which the selected categories affect population, land price, and passengers by using the multiple regression method. This research can aid the further development of land and building usage in the future.

==Keywords==

Railway and subway stations ; Multiple regression method ; Population ; Land price ; Passengers

==1. Introduction==

===1.1. Background===

The rate of increase in the number of cars in the city is much higher than the development of road traffic infrastructure with the development of social economy and rapidly increasing urbanization. Urban rail traffic construction has thus entered into a high-speed development stage.

The road transportation system is burdened by the increasing number of urban cars. Vehicle emissions, such as carbon monoxide, hydrocarbons, nitrogen oxides, and photochemical products are harmful to residents׳ health. Buildings replace the ecological environment. As a result, traffic congestion, lack of parking space, environmental degradation, and so on, have become common in cities.

Almost all countries have prioritized the development of public transport, thereby creating experiences worth sharing. The United States and Europe have focused on automobiles to the detriment of public transport. The excessive pursuit of individual actions and the correct direction of sustainable development are lessons worth using as reference. Having urban rail transit as the backbone of urban public transportation and regular public transportation as an auxiliary system is generally considered the only feasible means of solving the urban traffic problem. Urban railways and subways are safe, comfortable, convenient, fast, effective, and environment-friendly urban infrastructure systems and reduce problems faced by many cities, such as shortages in land resources, traffic congestions, and air pollution. The quantified effects of rail transit development on land and building usage in adjacent areas must be examined to establish the effects of rail transit development on the population, land price, and the usage of adjacent land.

The theory of land use has been extended to address the dynamics of urban growth and decline. This extension has fundamental implications. First, the determinants of urban spatial structure vary between static and dynamic models. For example, in a static mono-centric model, lot sizes increase and densities decrease as distance from employment increases, because equilibrium land rents decline to offset the rising cost of commuting. However, in a dynamic model with durable housing and myopic landowners, urban development is an incremental process, where densities depend solely on the economic conditions at the time of development. Densities may nevertheless decline with distance, because economic conditions change over time in particular ways (incomes increase or transportation costs fall), but land rent and selling prices and population densities may also rise with commuting distance.

An increasing number of buildings are being built as a result of the rebuilding of stations with the development of land readjustment and buildings along stations in Fukuoka city. Areas along railway tracks have thus changed rapidly. Recently, residential areas have gradually appeared in the center of the city, despite the tendency of resident to transfer to the suburbs. Furthermore, modern society still develops around public traffic; peripheral urban development centers on train stations are thus expected to become increasingly important to city planning centered on public traffic in an environmental society. With the development of large-scale retailers and specialty stores, the shutter streets of shopping streets have become increasingly serious. A shutter street is a street with various branch shops and closed-down shops or offices. Such areas used to form a busy shopping district with the taste of a “commercial town”, but with the shops closing down, the area is now mostly residential.

The Fukuoka-based Kyushu Economic Research Center has reported that more people left the Kyushu-Okinawa region than moved into the area in 2012, resulting in the first excess outflow of population in two years. The 4860-person drop was attributed to a decline in the number of people moving in from regions affected by the Great East Japan Earthquake and nuclear power plant incident, coupled with an increase in the number of people moving back to those regions in connection with disaster reconstruction efforts. Aside from Fukuoka and Okinawa, every prefecture in the region experienced excess outflow, led by Nagasaki and Kagoshima, with respective declines of 4906 and 3599 persons. Fukuoka experienced an inflow of 9221 persons, with over 60% of the new population originating from the Kyushu region (Source: Nishinippon News).

===1.2. Previous studies===

Over the years, researchers have introduced a large number of developments in railway and subway stations, and have examined the relationship between urban land and stations as well as the varying importance of stations in a city based on one criterion or another. The studies described in this section have proven highly valuable in the analysis and understanding of the roles played by railway and subway stations in the development of a city, such as the theory of rail transit and its development, effects on land market, the influence on house price, the developing situation in terms of land and building usage, analysis of changes in land price and house price, and so on. Given the three main elements supporting this paper, the previous studies are shown in turn.

Sustainable development and livable communities represent broad visionary ideas in contemporary urban planning. However, attempts to implement these popular visions face a host of conflicts. The future of land use planning may well depend on how to copes with these conflicts ([[#bib15|Godschalk, 2004]] ). [[#bib43|Wang and Yu (2012)]]  examined the characteristics of urban landscapes following a temporal-spatial pattern that provides a reference for the characteristics of land use in the temporal-spatial pattern. The influence of transit-oriented development (TOD) on the San Diego, CA condominium market indicates that TOD has a synergistic value greater than the sum of its parts. This result also implies a healthy demand for additional TOD housing in San Diego ([[#bib11|Duncan, 2011]] ). [[#bib25|Kim (2010)]]  examined how land use planning and regulation may affect regional economic prosperity by reviewing the relevant literature. [[#bib17|Handy (2005)]]  noted a connection between transportation and land use. [[#bib23|Kestens et al. (2006)]]  introduced household-level data into hedonic models to measure the heterogeneity of implicit prices based on the household type, age, educational attainment, income, and previous tenure status of the buyers. Hess evaluated the influence on the distance from Buffalo, New York to light railway stations on housing price. After using price models to assess the housing prices of 14 light rail spots, Hess concluded that for each foot away from the light rail station, the average value of the house increased by $2.31 or $0.99 according to linear distance calculations ([[#bib19|Hess, 2007]] ). [[#bib7|Cohen and Paul (2007)]]  evaluated the effects of enhanced transportation systems on property values. [[#bib39|Ryan (2011)]]  reviewed the relationships among transportation facilities, highways, heavy rail, and light rail transit systems and property values. There are also some researchers who have studied the similar subjects with this paper [[#bib29|Luca (1996)]] , [[#bib30|Luca and Tejo (2001)]] , [[#bib31|Makrí and Folkesson (1999)]] , [[#bib36|Porta et al. (2005)]] , [[#bib37|Ratti (2004]] ).      

The high power and reliability of metabonomic data analysis using nuclear magnetic resonance (NMR) spectroscopy together with chemometric techniques to explore and predict the toxic effects in rats has been demonstrated ([[#bib4|Beckonert et al., 2003]] ). [[#bib42|Timm et al. (2004)]]  explored a possibilistic fuzzy clustering approach that lacks the severe drawback of the conventional approach, namely, that the objective function is truly minimized only if all cluster centers are identical. [[#bib27|Li and Yeh (2004)]]  analyzed urban expansion and the spatial restructuring of land use patterns in the Pearl River Delta in south China using remote sensing and GIS. Based on the methodology and results, [[#bib12|Farber and Yeates (2006)]]  compared the results of the application of a number of spatial multivariate models to two “global” models in a hedonic house price context. [[#bib5|Bitter and Mulligan (2007)]]  compared two approaches (spatial expansion method and geographically weighted regression) to examine spatial heterogeneity in housing prices in the Tucson, Arizona housing market. Yan Zhuang analyzed the effects of subway stations on commercial land values, using linear regression and regression analysis to identify the significant effects of subway stations on commercial land price, thereby establishing a model of the spatial relations between land prices and distance ([[#bib49|Zhuang and Zhuang, 2007]] ). An innovative LUR method implemented in a GIS environment that reflects both temporal and spatial variability and considers the role of meteorology is presented by [[#bib41|Su et al. (2008)]] . [[#bib46|Yano (2006)]]  has done a research by GIS on Changes of Interprise Location among Stations in Kyoto which has been useful reference for this paper. There are also some research methods which have inspired key points of this paper ([[#bib2|Babalik-Sutcliffe (2002)]] , [[#bib3|Bafna (2003)]] , [[#bib18|Handy and Niemeier (1997)]] , [[#bib20|Honda (1977)]] , [[#bib21|Iwatani (2002)]] , [[#bib24|Kim et al. (2003)]] , [[#bib47|Yasuhiko (2000)]]  and [[#bib50|Zhu and Wang (2005)]] ).      

[[#bib33|Naceur (2013)]]  conducted a case study of Bouakal, Batna focusing on how modifications of urban space in informal settlements influence residents׳ quality of life, and examined the effects of urban improvement in Bouakal. This article showed the influence on urban development and the relationship between the urban space and the settlemetns. [[#bib26|Knight and Trygg (1977)]]  have drawn the impacts of rapid transit systems on urban development. It is concluded that rapid transit can have substantial growth-focusing impacts. [[#bib16|Golub et al. (2012)]]  showed that proximity to light rail transit stations positively affects property values and land price. Urban growth affects land price. In a static context, land price is proportional to the price of land at the boundary of an urban area, which is equal to the value of agricultural land rent. In a dynamic context, neither property holds. Previous research has shown that land price has four additive components: the value of agricultural land rent, the cost of conversion, the value of accessibility, and the value of expected future rent increases. In a dynamic context, an efficient land market naturally produces a gap between the price of lands at the boundary (minus conversion cost) ([[#bib6|Capozza and Helsley, 1989]] ). [[#bib38|Rosenthal and Helsley (1994)]]  demonstrated a new methodology for recovering vacant urban land prices in developed areas. They also estimated the price of vacant land based on properties that had been sold and redeveloped. The price of developed land was estimated on the basis of properties that had been sold and not redeveloped. Various tests overwhelmingly supported the hypothesis that housing is redeveloped when the price of vacant land exceeds the price of land in its current use. This result supports various theoretical models of urban spatial growth and indicates that one can recover vacant urban land prices by examining redeveloped properties. One case discusses the distribution of land prices in Jakarta using information provided on a neighborhood basis by experienced real estate brokers. Analysis of the data from Jakarta shows the relative importance of infrastructural provision and tenure in determining land prices ([[#bib10|Dowall and Leaf, 1991]] ). [[#bib28|Lin and Evans (2000)]]  analyzed the relationship between land price and plot size when the plots are small. In another case analysis on the land price, the price of residential land in Mexico declined significantly in real terms during the 1980s. Land prices appeared to follow a cyclical trend that tracked the macro-economic performance of Mexico ([[#bib44|Ward et al., 1993]] ). With regard to the research on changes in land price, [[#bib35|Ping and Chen (2004)]]  establish a data set about 35 metropolitan areas in China based on the “Survey of China׳s real estate industry”(1999–2002), and discuss the effects of government policy on land price control and of credit supply on the development of real estate industry. They have recently found that the effect of decreasing land price is greater than the effect of increasing house price on real estate investment, and that foreign capital is an active factor in promoting the expansion of China׳s real estate industry. A significant relationship exists among the dynamics of house prices, structural costs, and land prices, which has resulted in the first constant-quality price and quantity indexes for the aggregate stock of residential land in the United States. In a range of applications, we show that these series can shed light on trends, fluctuations, and regional variations in the price of housing ([[#bib9|Davis and Heathcote, 2007]] ). [[#bib48|Zeng and Zhang (2006)]]  explored the relationship between land price and housing price using positive analysis and econometrics analysis. The result of this study implies that short term, interactive effects exist in Wuhan, in which land price affects house price more than house price affects land price; however, no significant casual relationship exists between them in the long term. Moreover, [[#bib40|Song and Gao (2007)]]  concluded that land price is independent of housing price; moreover, although land price Granger causes housing price in the short run, housing price and land price interact as both cause and effect in the long run. Thus, long-term solutions to curbing housing price include controlling land price, rationalizing land supply, and obtaining an effective house construction plan. [[#bib8|Damm et al. (2007)]] , [[#bib14|Ghebreegziabiher et al. (2007)]] , [[#bib32|Morancho (2003)]] , and [[#bib34|Paez et al. (2001)]]  have done a similar research with this paper which are useful references.      

Based on the abovementioned research background and previous studies, the following points can be made:
* Based on the review of literature, studies have mostly focused on the analysis of urban railway stations and their effects on the increasing value of real estate in surrounding areas. By contrast, studies on the effects of these urban railway stations on the development of land use, the increase and decrease on the size of the population, the rise and fall of land prices, and the increase and decrease in passengers per day are limited;
* Almost all the existing studies are focused on industrial real estate and residential price, and less involved in the relationship between urban development and land price. Thus, this study focuses on the effect of annual change in land use on land price using Fukuoka, Japan as the study area;
* Previous studies have conducted a demonstrational analysis of the combined effects of time and space in a limited manner. Thus, this study analyzes the effects of the changes in land use on population, land price, and passengers by using multiple regression analysis to provide a reference for the development of real estate and urban planning, and to create a large comprehensive transportation system in different directions and construction periods.

===1.3. Objectives and framework===

According to the background and the previous studies described above, by using relevant data on the development status of land use and building usages, a series of analyses are launched. Changes in the land use within and outside the station zones, as well as the annual developing trend, are then interpreted. The developing trends in the Fukuoka station zones are subsequently determined. Using the data on the land use and building usages around the railway stations and subways in Fukuoka, this study selects 68 stations and eight lines as the research targets, which are compared and analyzed based on the distribution of land use by extracting the POSMAP data on land use with the GIS. In addition, these data obtained over five years are classified into two groups, that is, 0 m–400 m radius group (Step 1) and 400 m–800 m radius group (Step 2).

Based on the abovementioned backgrounds and the previous analysis, the effects of the changes in the surrounding regions on population, land price, and number of passengers, as well as the effect of station zones on annual changes need to be investigated. Thus, this study aims to determine the effects of land use on annual changes by analyzing the different distances of the surrounding regions in station zones. Moreover, the developing trend is analyzed using quantitative analysis, that is, by using the multiple regression algorithm.

This paper consists of five parts:
* Part 1 focuses on the general introduction including the research background, previous studies and research objectives;
* Part 2 contains the research objects, that is, the selection of the target stations, and an overview of the data on hand;
* Part 3 shows the analysis of the development status of and annual changes in the land use and building usage of the 68 target stations and the data obtained from the Geographic Information System;
* Part 4 discusses the use of multiple regression analysis to determine the effects of the changes in land use and building usage on population, land price, and number of passengers. In addition, the relationships among population, land price, number of passengers, and changes in land use and building usages are determined;
* Part 5 is a summary of this paper, which includes the analysis results and possible solutions for these problems.

==2. Research objects==

===2.1. Target city===

Fukuoka City, which is the object of this research, has eight lines and 68 stations in total. These lines are the JR Kagoshima Line, the JR Kashii Line, the JR Chikuhi Line, the Nishitetsu Omuta Line, the Nishitetsu Kaizuka Line, and three lines from the Fukuoka Municipal Subway, which are shown in [[#f0005|Figure 1]] . The annual developments of these 68 stations and eight lines in Fukuoka City are analyzed.

<span id='f0005'></span>

{| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; max-width: 100%;" 
|-
|


[[Image:draft_Content_379113966-1-s2.0-S2095263514000053-gr1.jpg|center|508px|Outline of object routes and stations.]]


|-
| <span style="text-align: center; font-size: 75%;">

Figure 1.

Outline of object routes and stations.

</span>
|}

===2.2. Overview of data===

This study uses the data on land use and building usage obtained using the POSMAP from five years (1985, 1993, 1998, 2003, and 2008) for each station, as well as uses the GIS to extract data on land use and building use. The data on population and number of households, which are obtained from the census of the population of Fukuoka City for 1985, 1990, 1995, 2000, 2005, and 2010, and are in units of chome (the chome is a special street unit in Japan; the “cho” in Japanese is a unit for streets; and the “me” in Japanese indicates the number; hence, 1 Chome means the first avenue). The number of passengers are obtained using the statistical books of Fukuoka and the data for over 36 years, that is, from 1975 to 2011, for all routes of JR and the Japanese western railways ([[#bib13|Fukuoka City Statistics Book, 2000]] ). The data on the Fukuoka City Subway were collected from its opening year in 1981 until 2011, a total of 30 years ([[#bib1|A Research Report Concerning on Urban Structure of Fukuoka City, 2005]] ). The data on the survey for prefectural land price and public reviewing land prices, which are obtained from the National Land Numeric Information Download Service of the Ministry of Land, Infrastructure, Transport, and Tourism, were collected from 1983 to 2011, a total of 28 years. The data on architectural confirmation are the building confirmation data for Fukuoka City collected from 1992 until 2004, a total of 13 years. These data are all available and integrated using GIS.      

===2.3. Distributions of station zones in 1985 and 2005===

====2.3.1. Distribution of station zones in 1985====

[[#f0010|Figure 2]]  shows that the distribution of the station zones of Fukuoka City in 1985, which gradually overlapped with the Y-shaped urban structure of Fukuoka, has clearly changed. Some areas in the station zones had fault zones. In addition, the southwestern part of Fukuoka City had a large number of residential areas which were moved to outside districts of the station zones, which were mainly centered on buses and cars.

<span id='f0010'></span>

{| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; max-width: 100%;" 
|-
|


[[Image:draft_Content_379113966-1-s2.0-S2095263514000053-gr2.jpg|center|508px|Distributions of the station zones in the year 1985.]]


|-
| <span style="text-align: center; font-size: 75%;">

Figure 2.

Distributions of the station zones in the year 1985.

</span>
|}

====2.3.2. Distributions of station zones in the year 2005====

[[#f0015|Figure 3]]  shows the distribution of the station zones in Fukuoka City in 2005, which was based on the Y-shaped urban structure of Fukuoka City. With the opening of the Fukuoka Subway Nanakuma Line in 2005, the acreage of station zones has rapidly increased. Furthermore, the fault zones in 1985, which had always been broken, have become intact again since 2003.

<span id='f0015'></span>

{| style="text-align: center; border: 1px solid #BBB; margin: 1em auto; max-width: 100%;" 
|-
|


[[Image:draft_Content_379113966-1-s2.0-S2095263514000053-gr3.jpg|center|508px|Distributions of the station zones in the year 2005.]]


|-
| <span style="text-align: center; font-size: 75%;">

Figure 3.

Distributions of the station zones in the year 2005.

</span>
|}

==3. Quantitative analysis==

===3.1. Land use===

First, with regard to quantitative analysis, [[#bib45|Yang et al. (2012)]]  analyzed functional areas such as urban centers, with emphasis on the contradictions involved in intensive land use. In their study, they considered three aspects of intensive use, that is, buildings, lands, and traffic, as well as multiple evaluation factors based on quantitative research methods. The methods and patterns that they used provided us some hints for our study. We also use a quantitative research method to obtain data on the average sizes and percentages of the areas within the 0 m–400 m and 400 m–800 m radii of the stations in 5 years, which are determined using land use and building usage parameters. Furthermore, [[#bib45|Yang et al. (2012)]]  conducted a horizontal research by means of quantitative and comparative studies on individual factors, developed the evaluation model for intensive land use in urban centers, and analyzed the driving forces of intensive land use from buildings, land use, roads, and so forth.      

[[#t0005|Table 1]]  shows the average sizes and percentages of the areas determined using the parameters for land use and building usage, which are within the 0 m–400 m radius of the stations for five years. [[#t0005|Table 1]]  also shows that the 13 parameters for the land use, namely, commerce, housing, government, industry, transportation, parks and green land, used space, unused space, agriculture, forest, water, roads, and others. To determine the annual changes in land use, the average size and percentage of each area for five years are calculated. The average size is obtained by using GIS. The original data on the 68 stations for 1985, 1993, 1998, 2003, and 2008 are used as input for the GIS. The radius zone of 0 m–400 m is localized and a buffer analysis on this radius zone is this conducted. The size of the 13 parameters of land use can then be extracted. Furthermore, the data on the commerce for the 68 stations are added and the average value is obtained. The average values for the other parameters are obtained in the same manner.

<span id='t0005'></span>

{| class="wikitable" style="min-width: 60%;margin-left: auto; margin-right: auto;"
|+

Table 1.

Average size and percentage of the areas distinguished by the factors of land use and building usage within a radius of 0 m–400 m of the stations.

|-

! rowspan="2" | Usages
! rowspan="2" | Factors
! colspan="2" | 1985
! colspan="2" | 1993
! colspan="2" | 1998
! colspan="2" | 2003
! colspan="2" | 2008
|-

! Size
! %
! Size
! %
! Size
! %
! Size
! %
! Size
! %
|-

| rowspan="13" | Land use
| Commerce
| 138,022
| 9
| 127,835
| 9
| 131,513
| 9
| 123,401
| 8
| 145,618
| 11
|-

| House
| 387,788
| 26
| 433,517
| 29
| 461,733
| 31
| 468,085
| 32
| 373,923
| 27
|-

| Government
| 160,896
| 11
| 163,573
| 11
| 154,044
| 10
| 157,325
| 11
| 150,949
| 11
|-

| Industry
| 23,663
| 2
| 24,065
| 2
| 20,293
| 1
| 15,647
| 1
| 17,049
| 1
|-

| Transportation
| 66,626
| 4
| 67,459
| 4
| 60,880
| 4
| 60,452
| 4
| 67,208
| 5
|-

| Park and green land
| 91,902
| 6
| 78,408
| 5
| 94,134
| 6
| 92,237
| 6
| 82,513
| 6
|-

| Used space
| 64,180
| 4
| 112,112
| 7
| 88,970
| 6
| 90,437
| 6
| 99,622
| 7
|-

| Unused space
| 66,017
| 4
| 2228
| 0
| 15,454
| 1
| 18,876
| 1
| 12,656
| 1
|-

| Agriculture
| 96,872
| 6
| 71,196
| 5
| 61,362
| 4
| 51,183
| 3
| 53,343
| 4
|-

| Forest
| 75,153
| 5
| 65,642
| 4
| 57,433
| 4
| 54,116
| 4
| 51,159
| 4
|-

| Water
| 62,037
| 4
| 58,929
| 4
| 58,890
| 4
| 50,248
| 3
| 49,594
| 4
|-

| Road
| 218,417
| 15
| 229,623
| 15
| 262,692
| 18
| 268,087
| 18
| 231,944
| 17
|-

| Others
| 41,223
| 3
| 65,286
| 4
| 24,509
| 2
| 29,445
| 2
| 24,599
| 2
|-

| colspan="12" | 
|-

| rowspan="8" | Building usage
| Business and hotel
| 173,136
| 19
| 190,420
| 19
| 58,890
| 5
| 50,248
| 4
| 56,587
| 4
|-

| Entertainment
| 74,919
| 8
| 95,591
| 10
| 262,692
| 21
| 268,087
| 20
| 298,180
| 21
|-

| Detached house
| 253,115
| 27
| 169,521
| 17
| 24,509
| 2
| 29,445
| 2
| 26,759
| 2
|-

| Condominium
| 197,625
| 21
| 319,214
| 32
| 174,314
| 14
| 189,675
| 14
| 175,360
| 13
|-

| Government
| 177,849
| 19
| 179,512
| 18
| 80,311
| 6
| 90,381
| 7
| 92,568
| 7
|-

| Transportation
| 25,238
| 3
| 22,347
| 2
| 154,533
| 12
| 150,489
| 11
| 164,435
| 12
|-

| Industry
| 25,789
| 3
| 15,386
| 2
| 376,890
| 30
| 404,691
| 31
| 443,681
| 32
|-

| Others
| 7080
| 1
| 4986
| 1
| 133,449
| 11
| 141,532
| 11
| 131,265
| 9
|}

[[#t0005|Table 1]]  shows that the percentages for commerce, government, transportation, park and green land, forest, and water have not change significantly over the five years. The reason is, in recent years, the development of Fukuoka, Japan has stabilizes and parameters, such as the government and the transportation, have not changed. The percentages for the house and the road continue to increase, whereas the percentages for the industry and the agriculture continue to decrease, resulting from the development of tertiary industries. With urban development, the numbers of buildings and houses for people to live in are increasing. At the same time, for convenience purposes, the number of roads, which are necessary, is also increasing. As tertiary industries continue their rapid development, the percentages for industry and agriculture have been gradually decreasing. Moreover, the percentages for used space, unused space, and others are always changing, sometimes increasing, and sometimes decreasing. However, a lack of significant difference in terms of percentage is observed between used space and unused space. Given that the uncertainty of the above parameters, thus, the observed changes for these parameter are likewise undefined. For example, a used space can also become unused when it is abandoned.      

From [[#t0005|Table 1]] , eight building usage influencing factors of are shown, which are abstracted by the same method as in land use data. The building usage factors are as follows: business and hotel, entertainment, detached house, condominium, government, transportation, and industry. The percentage of entertainment, transportation, and industry steadily increased from 1985 to 2008 because of the speeding up of the urbanization in Fukuoka and the people׳s basic needs. People have paid more attention to the development of entertainment and transportation. The percentage of business and hotel, detached house, condominium, and government has been decreasing, indicating that more buildings on housing are being converted into other functions. With the development in the tertiary industry, the percentage of entertainment has always been in increasing.      

===3.2. Building usage===

[[#t0010|Table 2]]  shows the average size and percentage of the areas distinguished by the land use and building usage factors within a radius of 400 m to 800 m of the stations every 5 years. This table shows 13 land use influencing factors, and the same method is used to analyze the average size of land use and building usage as in [[#t0005|Table 1]] .

<span id='t0010'></span>

{| class="wikitable" style="min-width: 60%;margin-left: auto; margin-right: auto;"
|+

Table 2.

Average size and percentage of the areas distinguished by the factors of land use and building usage within a radius of 400 m–800 m of the stations in every 5 years.

|-

! rowspan="2" | Usages
! rowspan="2" | Factors
! colspan="2" | 1985
! colspan="2" | 1993
! colspan="2" | 1998
! colspan="2" | 2003
! colspan="2" | 2008
|-

! Size
! %
! Size
! %
! Size
! %
! Size
! %
! Size
! %
|-

| rowspan="13" | Land use
| Commerce
| 196,445
| 8
| 183,934
| 8
| 198,669
| 8
| 191,734
| 8
| 141,297
| 7
|-

| House
| 562,102
| 24
| 640,781
| 27
| 723,468
| 29
| 694,969
| 29
| 514,814
| 26
|-

| Government
| 211,252
| 9
| 230,662
| 10
| 217,120
| 9
| 221,171
| 9
| 155,946
| 8
|-

| Industry
| 37,701
| 2
| 42,909
| 2
| 38,384
| 2
| 30,932
| 1
| 29,765
| 1
|-

| Transportation
| 55,977
| 2
| 58,321
| 2
| 63,177
| 3
| 68,655
| 3
| 76,671
| 4
|-

| Park and green land
| 147,528
| 6
| 120,689
| 5
| 162,815
| 7
| 163,203
| 7
| 154,182
| 8
|-

| Used space
| 132,693
| 6
| 174,438
| 7
| 135,821
| 6
| 127,597
| 5
| 115,249
| 6
|-

| Unused space
| 123,947
| 5
| 2555
| 0
| 49,212
| 2
| 53,842
| 2
| 55,434
| 3
|-

| Agriculture
| 153,718
| 6
| 113,249
| 5
| 100,183
| 4
| 79,578
| 3
| 59,854
| 3
|-

| Forest
| 219,711
| 9
| 245,084
| 10
| 220,756
| 9
| 202,234
| 9
| 213,979
| 11
|-

| Water
| 101,186
| 4
| 94,091
| 4
| 100,454
| 4
| 88,290
| 4
| 86,325
| 4
|-

| Road
| 396,461
| 17
| 358,217
| 15
| 407,518
| 17
| 402,862
| 17
| 335,579
| 17
|-

| Others
| 44,648
| 2
| 102,679
| 4
| 48,679
| 2
| 48,419
| 2
| 48,522
| 2
|-

| colspan="12" | 
|-

| rowspan="8" | Building usage
| Business and hotel
| 101,186
| 6
| 94,091
| 5
| 100,454
| 5
| 88,290
| 5
| 96,712
| 5
|-

| Entertainment
| 396,461
| 23
| 358,217
| 19
| 407,518
| 22
| 402,862
| 21
| 465,223
| 23
|-

| Detached house
| 44,648
| 3
| 102,679
| 5
| 48,679
| 3
| 48,419
| 3
| 50,021
| 2
|-

| Condominium
| 236,774
| 14
| 250,076
| 13
| 229,640
| 12
| 246,101
| 13
| 253,345
| 12
|-

| Government
| 91,719
| 5
| 130,001
| 7
| 116,470
| 6
| 137,405
| 7
| 112,031
| 5
|-

| Transportation
| 358,538
| 21
| 244,262
| 13
| 242,053
| 13
| 217,374
| 11
| 300,188
| 15
|-

| Industry
| 305,666
| 18
| 476,485
| 25
| 576,586
| 31
| 601,798
| 31
| 605,233
| 29
|-

| Others
| 208,842
| 12
| 219,459
| 12
| 163,455
| 9
| 169,739
| 9
| 177,549
| 9
|}

In [[#t0010|Table 2]] , the percentages of commerce, government, water, and road stay at a stable situation, which shows these facilities at a constant state after a rapid period of development. The percentages of house, transportation, park, and green land also have steadily increased, led by gradually increases in houses at the radius of 400 m–800 m. House and transportation has a close relationship, thus, both have been simultaneously developing. Notably, industry and agriculture are in a decreasing trend in the radius of 400–800 m. Three special factors (i.e., used space, unused space, and forest) have been in a fluctuating trend from 1985 to 2008. In the building usage of [[#t0010|Table 2]] , business and hotel, as well as detached house, remain at a steady situation. The percentage of industry has been increasing, whereas the others are decreasing. The percentages of entertainment, condominium, government, and transportation are also increasing with the fluctuation for five years.      

===3.3. Population, land price, and passengers===

[[#t0015|Table 3]]  shows the total population within the radius of 0 m–400 m and 400 m–800 m, average land price, and the total passengers per day of 68 stations from 1985–2008. The land price in these periods fluctuated, reaching the highest point in 1993 and the lowest in 1985. For population both in the radius of 0 m to 400 m and 400 m to 800 m, the population is always increasing because of the influence of development around stations. The passengers were also fluctuating.

<span id='t0015'></span>

{| class="wikitable" style="min-width: 60%;margin-left: auto; margin-right: auto;"
|+

Table 3.

Total population within the radius of 0 m–400 m and 400 m–800 m, average land price and total passengers per day of the stations in every 5 years.

|-

! colspan="2" | Factors
! Unit
! 1985
! 1993
! 1998
! 2003
! 2008
|-

| rowspan="2" | Population
| 0 m–400 m
| Person
| 716,897
| 758,989
| 766,953
| 793,541
| 839,899
|-

| 400 m–800 m
| Person
| 1,017,245
| 1,094,192
| 1,125,024
| 1,176,516
| 1,252,527
|-

| colspan="2" | Land price
| Japanese Yen
| 255,304
| 869,364
| 404,673
| 289,023
| 339,232
|-

| colspan="2" | Passengers
| Person/day
| 1,844,817
| 2,711,356
| 2,757,008
| 2,565,824
| 2,633,687
|}

==4. Regression algorithm==

The regression algorithm is used in this research, which is a statistical technique to estimate the relationships among variables. It includes many techniques to model and analyze several variables, focusing on the relationship between a dependent variable and one or more independent variables. Regression analysis is the most popular analysis mode in the multivariate analytic method, which is the representative of the “predictive and forecasting mode”. It is also used to understand which among the independent variables are related to the dependent variable and explore the forms of these relationships. Specifically, regression analysis helps in understanding how the typical value of the dependent variable changes when any one of the independent variables is varied, while the other independent variables are fixed. Regression analysis more commonly estimates the conditional expectation of the dependent variable given the independent variables, that is, the average value of the dependent variable when the independent variables are fixed. The focus is less commonly on a quantile or other location parameters of the conditional distribution of the dependent variable given the independent variables. In all cases, the estimation target is a function of the independent variables called the regression function. In regression analysis, characterizing the variation of the dependent variable around the regression function is of interest, which can be described by a probability distribution.

If two or more independent variables exist, this condition is called multiple regressions. A phenomenon is often associated with a plurality of factors; based on the optimum combination of the plurality of independent variables, the dependent variable can be predicted or estimated, which is more effective and more realistic than forecasting or estimating with only one variable.

This research uses the variable selection method of multiple regression analysis, that is, the method of reducing variables and increasing variables to choose relevant variables. Afterwards, by standardizing the partial regression coefficients, the influence degree of annual changes in land use in the surrounding regions of railway and subway stations can be clearly defined. The variable whose absolute value of partial regression coefficients is more than 1 after being standardized from the angle of its relationship with multi-collinearity is considered. If the absolute value of partial regression coefficients of the received variable is unsuitable for further analysis, all the variables will be recombined repeatedly until an appropriate variable can be obtained.

===4.1. Data outline===

Based on the above analysis, using a GIS to extract the POSMAP data on land use and building usage of station zones, as well as further grouping these data by cluster analysis, this study organizes the data according to different times, applications, and distances by multiple regression analysis. The 13 variables used in terms of land use are “Commerce”, “House”, “Government and Education”, “Transportation”, “Park and Green Land”, “Used Land”, “Unused Land”, “Agriculture”, “Forest”, “Water”, “Road”, and “Others”. The eight variables used in terms of building usage are “Business and Hotel”, “Entertainment”, “Detached House”, “Condominium”, “Government and Education”, “Transportation”, “Industry”, and “Others”. The multiple regression analysis can then be processed with the above variables.

===4.2. Formulas of the multiple regression analysis===

The process for the multiple regression analysis is as follows. First, the paper uses the variable selecting method of the multiple regression analysis to reduce or increase variables. According to the factors and the explanatory variables, the multiple regression analysis can be then be performed as follows:

{| class="formulaSCP" style="width: 100%; text-align: center;" 
|-
| 
{| style="text-align: center; margin:auto;" 
|-
| <math>f=a_0+a_1x_1+a_2x_2+...+a_nx_n</math>
|}
| style="width: 5px;text-align: right;white-space: nowrap;" | 
|}

where ''f  ''  is the theoretical value, <math display="inline">a_n</math>  is the partial regression coefficient, and <math display="inline">x_n</math>  is the explanatory variable.      

In the calculation process of multiple regression algorithm, the ''t''  and ''P''  values are the basic values, and also usually used as reference values. Based on the ''t''  and ''P''  values, the process restructures the variables and carries out the multiple regression analysis by ignoring the variable ''x'' .      

Several conditions are used: the ''t''  and ''P''  values are positive; the relevant variable ''r'' , the partial regression coefficent ''a'' , and the symbol are consistent; and the absolute value of the standardized partial regression coefficients is ˗1. When the above conditions are all met, the remaining variables become the basic factors for the regression analysis. Grasping the influence degree of each basic factor quantitatively from the standardized partial coefficients is also the final step for the process of the multiple regression method.      

===4.3. Variations of influence degree on time and distance===

====4.3.1. Influence degree on population====

Comparing Step 1 with Step 2, more variables are selected in Step 1. Concerning the multiple regression analysis on the population, a trend is observed wherein variables related to land use are frequently used in the multiple regression analysis. From [[#t0020|Table 4]] , the influence degree obtained from variables on land use tend to be negative, whereas the influence degree obtained from variables on building usage tend to increase. The reason why more variables are selected in Step 1 is because the land use categories around the station have been definite. In Step 2, a range of 400 m–800 m radius from the center of the station is abstracted, which gives it a doughnut shape. Land use also becomes totally different depending on directions, which is the impact of the “geographical orientation”.      

From the angle of different years and different usage, the influence of “apartment” in Step 1 increases every year. The large development of mansions around the center of each station has a great effect on the population size. Compared with Step 1, the influence degree of “apartment” in Step 2 increases, and the floor area of the apartment within the 400 m–800 m radius has a large proportion. “Business and Hotel” tends to be selected in Step 1, and the value of the influence degree is negative. This result shows that the location of business facilities has a negative impact on the population within a 400 m radius range. The influence degree on “road” in Step 1 has also obviously increased. Similar to road development, the population also had an increasing trend. The positive influence of road development on city planning is hence clear.

In relation to land use, most variables other than “house” and “road” are almost negative. The selected variables that have negative values include “government”, “transportation and warehouse”, “used space”, and “agriculture”. These variables, especially within a radius of 400 m, have a negative impact on the increase in the population size.

According to the data calculation and analysis, the regression equation and every variable coefficient have finished the significance test. The adjusted R2 value also fits well. The regression coefficients are shown in [[#t0020|Table 4]] . The regression equation shows that the growth of the population near the railway stations and subway stations greatly increased. However, the population density far from the railway and subway stations only gradually increased in recent years. This condition is caused by the prices of housing and land adjacent to the railway and subway stations being higher than the ones far away. More people have thus begun to live far from the rail transit sites in recent years. These areas are relatively quiet and have shopping malls that are convenient for people׳s everyday life. The closer the distance is, the higher the population density and land price become. Some peripheral areas around the railway and subway stations have not been developed well and utilized enough. The commerce and service are not as high as in other stations. Hence, these areas have lower population densities despite their proximity to the stations.      

According to the regression equation, the variables that have an effect on the population density are not so complex.

This paper summarizes the following two points:
* In general, the distance far from the railway and subway stations has a relatively lower population density.
* Commerce, school, and green land have a positive impact on changes in the population density. If a shopping mall, a school, or both are found near a station in a less developed region, then this region must be developing gradually, which is undoubtedly a positive impact.

====4.3.2. Influence degree on land price====

According to the results from the multiple regressions analysis on population ([[#t0020|Table 4]] ) and comparing Step 1 with Step 2, no obvious difference on the number of selected variables can be found. However, the corrected determination coefficient of Step 1 is higher, which is favorable as a regression equation. The multiple regression analysis on land price has frequently used more variables related on building usage. The influence degree of variables related to land use is also negative, whereas the variables related to the building usage seem to be positive. The reason why the corrected determination coefficient of the regression equation becomes higher in Step 1, which is almost the same as the multiple regression analysis on the population, is that clear and definite categories exist on land use around stations. In Step 2, as the different direction, land use also becomes very complicated and tolerance in variation exists, which makes it a lower determination coefficient.

<span id='t0020'></span>

{| class="wikitable" style="min-width: 60%;margin-left: auto; margin-right: auto;"
|+

Table 4.

Degree of influence on population by the factors of land use and building usage within the radius of 0 m–400 m and 400 m–800 m in every 5 years

|-

! rowspan="2" | Usages
! rowspan="2" | Factors
! colspan="5" | 0 m–400 m
! colspan="5" | 400 m–800 m
|-

! 1985
! 1993
! 1998
! 2003
! 2008
! 1985
! 1993
! 1998
! 2003
! 2008
|-

| rowspan="13" | Land use
| Commerce
| 
| 
| 
| 
| 
| −0.33
| 
| 
| 0.06
| 
|-

| House
| 0.43
| 0.47
| 0.44
| 0.19
| 0.23
| 
| 0.52
| 
| 0.49
| 0.51
|-

| Government
| 0.08
| −0.02
| 
| 
| 
| 
| 
| 
| 
| 0.12
|-

| Industry
| 
| 
| −0.12
| 
| 
| 
| 
| 
| 
| 
|-

| Transportation
| 
| −0.08
| 
| 
| 
| 0.09
| 
| 
| 
| 
|-

| Park and Green land
| 
| −0.08
| −0.34
| 
| 
| 
| −0.07
| 
| 
| 
|-

| Used space
| 
| 
| −0.21
| −0.10
| 
| 
| 
| 
| 
| 
|-

| Unused space
| 
| 
| 0.27
| 
| 0.25
| 
| 
| 
| 
| 
|-

| Agriculture
| −0.10
| −0.14
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Forest
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Water
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Road
| 
| 0.26
| 0.25
| 0.19
| 0.16
| 
| 
| 
| 
| 
|-

| Others
| 0.08
| 
| 
| 
| 
| 
| −0.10
| 
| 
| 0.11
|-

| colspan="12" | 
|-

| rowspan="8" | Building usage
| Business and Hotel
| 
| −0.32
| −0.37
| −0.22
| −0.20
| 
| 
| 
| 
| 
|-

| Entertainment
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Detached house
| 0.28
| 
| 
| 0.17
| 
| 0.52
| 
| 
| 0.61
| 
|-

| Condominium
| 0.52
| 0.48
| 0.68
| 0.81
| 
| 0.75
| 0.59
| 0.96
| 0.70
| 0.86
|-

| Government
| 
| 
| 
| 
| 
| 0.10
| 0.07
| 
| 0.03
| 
|-

| Transportation
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Industry
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Others
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| colspan="12" | 
|-

| rowspan="2" | Variables
| Chosen Variables
| 6
| 8
| 8
| 6
| 4
| 5
| 5
| 1
| 5
| 4
|-

| Revised R2
| 0.86
| 0.90
| 0.97
| 0.96
| 0.92
| 0.89
| 0.94
| 0.87
| 0.97
| 0.96
|}

From the angle of different years and different usage in Step 1, “detached house” in 1985 is chosen frequently, whereas “apartment” becomes widely in use after 1993, which has a larger impact on land price compared with “detached house”. Except for “population”, all the factors have a positive impact on land price. In recent years, the influence degree on the apartment has increased, which makes it an important factor for land price.

In relation to land use, the influence degree of housing is negative, indicating that the land prices fall with the spreading of land for housing. The reason seems to be that the expansion of a detached house leads to low land prices in the low-density residential area. Fewer variables are selected in Step 2, and the whole trend is mixed and difficult to be grasped. “House”, “business and hotel”, “commerce and entertainment”, “apartment”, “commercial and entertainment”, “business and lodging”, and “residential land” tend to be selected.

According to the calculating process of the regression equation and every variable coefficient, they have clearly finished the significance test. The adjusted R2 value also fits well. The regression coefficients are shown in [[#t0025|Table 5]] . The regression equation shows that the land price near the railway and subway stations are usually high. The closer to the rail transit site, the higher the land price becomes. Some peripheral areas around subway stations have not been developed well and utilized enough, and houses which have been built are almost apartments. The flourishing degrees of the surrounding commerce and service are not so high. The closer to this kind of districts, the lower the land price becomes.

<span id='t0025'></span>

{| class="wikitable" style="min-width: 60%;margin-left: auto; margin-right: auto;"
|+

Table 5.

Degree of influence on land price by the factors of land use and building usage within the radius of 0 m–400 m and 400 m–800 m in every 5 years.

|-

! rowspan="2" | Usages
! rowspan="2" | Factors
! colspan="5" | 0 m–400 m
! colspan="5" | 400 m–800 m
|-

! 1985
! 1993
! 1998
! 2003
! 2008
! 1985
! 1993
! 1998
! 2003
! 2008
|-

| rowspan="13" | Land use
| Commerce
| 
| 
| 
| 
| 
| 0.78
| 0.23
| 
| 
| 0.21
|-

| House
| −0.15
| −0.49
| −0.53
| −0.42
| −0.33
| −0.30
| 
| 
| −0.22
| 
|-

| Government
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Industry
| 
| 
| 
| 
| 
| 
| 
| 
| 0.28
| 
|-

| Transportation
| 
| 
| 
| 
| 
| −0.24
| 
| 
| 
| 
|-

| Park and Green land
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Used space
| 
| 
| 
| 
| 
| 0.18
| 
| 
| 
| 
|-

| Unused space
| −0.03
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Agriculture
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Forest
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Water
| 
| 
| 
| 
| 
| 
| −0.17
| 
| 
| 
|-

| Road
| 
| 
| −0.09
| −0.14
| 
| 
| 
| 0.23
| 
| 
|-

| Others
| 
| 
| 
| 
| 
| 
| 0.21
| 
| 
| 
|-

| colspan="12" | 
|-

| rowspan="8" | Building usage
| Business and hotel
| 0.86
| 0.55
| 0.77
| 0.87
| 0.78
| −0.39
| 
| 0.68
| 
| 
|-

| Entertainment
| 0.15
| 0.33
| 0.30
| 0.25
| 0.20
| 0.32
| 0.52
| 
| 0.43
| 0.49
|-

| Detached house
| 0.18
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Condominium
| 
| 0.33
| 0.43
| 0.42
| 0.46
| 
| 0.21
| 0.15
| 0.36
| 0.22
|-

| Government
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Transportation
| −0.11
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Industry
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Others
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| colspan="12" | 
|-

| rowspan="2" | Variables
| Chosen Variables
| 6
| 4
| 5
| 5
| 4
| 6
| 5
| 3
| 4
| 3
|-

| Revised R2
| 0.92
| 0.90
| 0.94
| 0.92
| 0.94
| 0.65
| 0.84
| 0.77
| 0.86
| 0.78
|}

During a period, the influence of the rail transit on land prices is in the form of a parabola. The distance that is too close to the rail transit sites will bring certain negative externalities, such as crowded traffic, noisy traffic flow, and crowded people. These factors are the reasons for such phenomenon. For instance, the land price that is higher than the Kashii JR Station can indicate the effect and impact of a vice center district in a city. Some stations׳ land prices tend to be low because of a lower orientation on land price than other stations. The land prices are thus lower than districts far away from this area.

According to the regression equation, the variables that can influence the land price are several and very complex. This paper summarizes the following points:
* The farther from the railway and subway stations, the lower the land price becomes. On the contrary, the closer to the railway and subway stations, the higher the land price becomes, which has been testified by the data, facts and investigation.
* A positive effect is observed from parks and schools. If a park, school, or both are being planned and finally found near one station in a less developed region, then this region must be growing gradually and becoming prosperous. If some buildings such as a shopping mall or school can be found in a developing region, it will be more prosperous than before.
* The relationship between the distance (from some very important stations, such as the Tenjin Subway Station) and land price appears to be a cubic curve. Being outside this kind of stations results in the rise, fall, and rise again in land prices. This condition indicates that the influence on the land price has gradually emerged from the beginning to the end of construction.

====4.3.3. Influence degree on the passengers====

Few variables are selected both in Steps 1 and 2, and the corrected determination coefficient is lower. Variables of land use and building usage have little impact on the number of passengers because of other factors.

Comparing Step 1 with Step 2, the corrected determination coefficient of Step 1 tends to be higher, and the selected variables have an effect on the number of passengers. In Step 2, the corrected determination coefficient tends to be low; thus, whether the selected variables have an impact on the number of passengers is uncertain.

In Step 1, “transportation and warehouse” tends to be selected, which indicates that the number of passengers is affected. “Commercial land” in Step 2 also tends to be selected, from which the number of passengers tends to increase with the commercial land increase.

According to the data calculation and analysis, the regression coefficients are shown in [[#t0030|Table 6]] . The regression equation shows that the number of passengers near the railway and subway stations appears to be much more than the suburban districts and the low-density stations. Similarly, the number of passengers of government and education-type stations also has a high percentage in relation with everyday use frequency. A rail transit, which is mainly consisted by commerce, green land, and educational facilities, usually has a high frequency on the number of passengers. Some stations that have not been well developed and utilized enough have a low number of passengers because of the relatively underdeveloped business. The effects of the development on commerce, business, and transportation can thus be seen in [[#t0030|Table 6]] .

<span id='t0030'></span>

{| class="wikitable" style="min-width: 60%;margin-left: auto; margin-right: auto;"
|+

Table 6.

Degree of influence on passengers by the factors of land use and building usage within the radius of 0–400 m and 400 m–800 m in every 5 years.

|-

! rowspan="2" | Usages
! rowspan="2" | Factors
! colspan="5" | 0 m–400 m
! colspan="5" | 400 m–800 m
|-

! 1985
! 1993
! 1998
! 2003
! 2008
! 1985
! 1993
! 1998
! 2003
! 2008
|-

| rowspan="13" | Land use
| Commerce
| 
| 
| 
| −0.40
| 
| 
| 0.40
| 0.76
| 0.56
| 0.62
|-

| House
| 
| 0.60
| 
| 
| 
| −0.94
| 
| 
| 
| 
|-

| Government
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Industry
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Transportation
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Park and Green land
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Used space
| 0.52
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Unused space
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Agriculture
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Forest
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Water
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Road
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Others
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| rowspan="8" | Building usage
| Business and hotel
| 
| 
| 
| 0.44
| 
| −0.75
| 
| 
| 
| 
|-

| Entertainment
| 0.42
| 0.48
| 
| 
| 0.49
| 
| 
| 
| 
| 
|-

| Detached house
| −0.40
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Condominium
| 
| 
| 
| 
| 
| 0.72
| 
| 0.72
| 
| 0.72
|-

| Government
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Transportation
| 0.51
| 0.30
| 0.96
| 
| 0.88
| 
| 
| 
| 
| 
|-

| Industry
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| Others
| 
| 
| 
| 
| 
| 
| 
| 
| 
| 
|-

| rowspan="2" | Variables
| Chosen variables
| 4
| 3
| 1
| 2
| 2
| 3
| 1
| 2
| 1
| 2
|-

| Revised R2
| 0.63
| 0.65
| 0.80
| 0.80
| 0.80
| 0.33
| 0.40
| 0.60
| 0.34
| 0.44
|}

==5. Conclusions==

===5.1. Results===

This research selected 68 stations as the research targets, compared and analyzed the distribution of 13 factors of land use and 8 factors of building usage, and indicated the relationship between the factors and the population, land price, and passengers around railway and subway stations every 5 years for a period of 24 years (i.e., 1985, 1993, 1998, 2003, and 2008) using the multiple regression analysis. By taking categories in land use and building usage as explanatory variables, this paper also shows changes in land use and building usage around the urban railway and subway stations and the influence on population, land price, and passengers in the quantitative expression method.

The integration development of the urban railway and subway with the land use can play a comprehensive effect. They do not only improve the efficiency of the land development and increase government revenue, but also bring a stable number of passengers for the railway and subway. The core issue of the integration development for the railway and subways with land development is a scientific analysis on the quantitative relationship between stations and land price. This research hence has an important theoretical and practical significance that can provide solutions for integration development.

Three interpretations are also put forward in this research as follows:
* First, the formation of large comprehensive transportation and communication systems of railway and subway stations has undoubtedly played an obvious added-value role for the surrounding land and housing prices. As improvements on the prosperity of commerce, business, and traffic conditions in this kind of regions exist, commerce and transportation has gradually become the main developing key in the surrounding areas where a new hot spot or vice commercial district is forming. The effect has hence been increasingly apparent in the future.
* From the point of time, space, and the distance, the land price and buildings that have a short distance to commercial center-type stations can increase further. In contrast, the railway and subway stations that are mainly unused center space have a relatively small increasing opportunity.
* For the municipal construction on added-value land, the government should reserve in advance so that it can be put into later development and construction. In some areas where commerce, business, industry and transportation have been good, the added-value effect on the land price and real estate is clearer. For this reason, real estate developers should attempt to choose such a location for priority development.

Consequently, this research method is not only applicable in Fukuoka City, but can also be generally applied to other cities or countries. We hope this research can be a good reference for the rapidly developing transportation systems around the world.

===5.2. Future work===

According to the related investigations of land use and building usage around railway and subway stations in the past two decades, one of the most important and consistent findings is analyzing and understanding the changes and development in land use and building usage using multiple regression analysis. This method concerns accessible places and more particularly deals with the problem of the influence degree of land use and building usage on population, land price, and passengers to serve a spatially and timely distributed demand.

This research is only the beginning of a series of other related research studies. It opens the door for a variety of studies that can investigate the effects of development on land use and building usage. Further development is required, and other factors such as calibrating the influence factors, integration of more influence factors, and testing their accumulative affects have to be considered to make the present research fully operational. A number of possible topics can be considered for further work in this research:
* Integrating the calculations with mathematical derivation would allow a wide range of users to study the influence degree on surrounding areas of changes and station development. Calculating the measures mathematically will obtain a more precise result on the influence degree.
* Incorporating more traffic tools such as buses, which is also an influencing element of traffic tools in the city, can enrich the research.
* The research also discusses the developing trend of stations in future work to make this theoretical idea more useful to explain the state of urban structure. The stations׳ developing trend will be divided into several different types. Specifically, the research will abstract parts of the stations that had obvious changes in different years (i.e., 1985, 1993, 1998, 2003 and 2008). During this period, the stations that had special and obvious changes will be chosen and further analyzed in later research work. For instance, what type was this station in 1985? After a 5- or 10-year development, what type has this station become? After a 20-year change, what type will this station change into? This topic can hopefully become an important guideline for other research and future urban development.
* One of the preconditions of the above research is to make a prediction on changes in land use and building usage in future. A useful, believable, and practicable method for future prediction is hence necessary.
* The research results for Fukuoka City cannot be used directly in other cities. Only some of these results can be learned and transferred in the proper outcome for use in other cities. In the near future, we plan to target more cities in more countries. We can thus provide more detailed and comprehensive results that can help urban planners in planning cities, especially the development of districts around subway and railway stations.

==References==

<ol style='list-style-type: none;margin-left: 0px;'><li><span id='bib1'></span>
[[#bib1|A Research Report Concerning on Urban Structure of Fukuoka City, 2005]] A Research Report Concerning on Urban Structure of Fukuoka City. Fukuoka Asian Urban Research Laboratory, 2005, Japan.</li>
<li><span id='bib2'></span>
[[#bib2|Babalik-Sutcliffe, 2002]] E. Babalik-Sutcliffe; Urban rail systems: analysis of the factors behind success; Transp. Rev., 22 (4) (2002), pp. 415–447</li>
<li><span id='bib3'></span>
[[#bib3|Bafna, 2003]] S. Bafna; An introduction of space syntax; Environ. Behav., 35 (2003), pp. 17–29</li>
<li><span id='bib4'></span>
[[#bib4|Beckonert et al., 2003]] O.E. Beckonert, M. Bollard, T. Ebbels, H.C. Keun, H. Antti, E. Holmes, J.K. Nicholson; NMR-based metabonomic toxicity classification: hierarchical cluster analysis and k-nearest-neighbour approaches; Anal. Chim. Acta, 490 (1) (2003), pp. 3–15</li>
<li><span id='bib5'></span>
[[#bib5|Bitter et al., 2007]] C. Bitter, G.F. Mulligan, S. Dall’erba; Incorporating spatial variation in housing attribute prices: a comparison of geographically weighted regression and the spatial expansion method; J. Geogr. Syst., 9 (1) (2007), pp. 7–27</li>
<li><span id='bib6'></span>
[[#bib6|Capozza and Helsley, 1989]] D.R. Capozza, R.W. Helsley; The fundamentals of land prices and urban growth; J. Urban Econ., 26 (3) (1989), pp. 295–306</li>
<li><span id='bib7'></span>
[[#bib7|Cohen and Paul, 2007]] J.P. Cohen, C.M. Paul; The impacts of transportation infrastructure on property values: a higher-order spatial econometrics approach; J. Reg. Sci., 47 (3) (2007), pp. 457–478</li>
<li><span id='bib8'></span>
[[#bib8|Damm et al., 2007]] D. Damm, S. Lerman, E. Lerner-Lamm,  ''et al.''; Response of urban real estate values in anticipation of Washington Metro; J. Transp. Econ. Policy, 14 (2007), pp. 315–336</li>
<li><span id='bib9'></span>
[[#bib9|Davis and Heathcote, 2007]] M.A. Davis, J. Heathcote; The price and quantity of residential land in the United States; J. Monet. Econ., 54 (8) (2007), pp. 2595–2620</li>
<li><span id='bib10'></span>
[[#bib10|Dowall and Leaf, 1991]] D.E. Dowall, M. Leaf; The price of land for housing in Jakarta; Urban Stud., 28 (5) (1991), pp. 707–722</li>
<li><span id='bib11'></span>
[[#bib11|Duncan, 2011]] M. Duncan; The impact of transit-oriented development on housing prices in San Diego, CA; Urban Stud., 48 (1) (2011), pp. 101–127</li>
<li><span id='bib12'></span>
[[#bib12|Farber and Yeates, 2006]] S. Farber, M. Yeates; A comparison of localized regression models in a hedonic house price context; Can. J. Reg. Sci., 29 (3) (2006), pp. 405–420</li>
<li><span id='bib13'></span>
[[#bib13|Fukuoka City Statistics Book, 2000]] Fukuoka City Statistics Book, Fukuoka City, Japan, 2000.</li>
<li><span id='bib14'></span>
[[#bib14|Ghebreegziabiher et al., 2007]] D. Ghebreegziabiher, P. Eric, R. Piet; The impact of railway stations on residential and commercial property value: a meta-analysis; J. Real Estate Finan Econ., 35 (2007), pp. 161–180</li>
<li><span id='bib15'></span>
[[#bib15|Godschalk, 2004]] D.R. Godschalk; Land use planning challenges: coping with conflicts in visions of sustainable development and livable communities; J. Am. Plan. Assoc., 70 (1) (2004), pp. 5–13</li>
<li><span id='bib16'></span>
[[#bib16|Golub et al., 2012]] A. Golub, S. Guhathakurta, B. Sollapuram; Spatial and temporal capitalization effects of light rail in phoenix from conception, planning, and construction to operation; J. Plan. Educ. Res., 32 (4) (2012), pp. 415–429</li>
<li><span id='bib17'></span>
[[#bib17|Handy, 2005]] S. Handy; Smart growth and the transportation-land use connection: what does the research tell us?; Int. Reg. Science Review, 28 (2) (2005), pp. 146–167</li>
<li><span id='bib18'></span>
[[#bib18|Handy and Niemeier, 1997]] S.L. Handy, D.A. Niemeier; Measuring accessibility: an exploration of issues and alternatives; Environ. Plan., 29 (7) (1997), pp. 1175–1194</li>
<li><span id='bib19'></span>
[[#bib19|Hess, 2007]] D.P. Hess; Impact of proximity to light rail rapid transit on station-area property values in Buffalo, New York; Urban Stud., 44 (5) (2007), pp. 1041–1068</li>
<li><span id='bib20'></span>
[[#bib20|Honda, 1977]] M. Honda; Shimada Im. Multivariate Statistical Technique for Management; Publishing Department of Sanno and Management University, Japan (1977)</li>
<li><span id='bib21'></span>
[[#bib21|Iwatani, 2002]] Iwatani M.A. GIS Fundamental Study on Location Tendency of Condominium and TOD Promotion Plan, Architectonics Research, Graduation Thesis Outline Collections, 2002.</li>
<li><span id='bib23'></span>
[[#bib23|Kestens et al., 2006]] Y. Kestens, M. Thériault, F. Des Rosiers; Heterogeneity in hedonic modeling of house prices: looking at buyers׳ household profiles; J. Geogr. Syst., 8 (1) (2006), pp. 61–96</li>
<li><span id='bib24'></span>
[[#bib24|Kim et al., 2003]] C.W. Kim, T.T. Phipps, L. Anselin; Measuring the benefits of air quality improvement: a spatial hedonic approach; J. Environ. Econ. Manag., 45 (1) (2003), pp. 24–39</li>
<li><span id='bib25'></span>
[[#bib25|Kim, 2010]] J.H. Kim; Linking land use planning and regulation to economic development: a literature review; J. Plan. Lit., 26 (1) (2010), pp. 35–47</li>
<li><span id='bib26'></span>
[[#bib26|Knight and Trygg, 1977]] R.L. Knight, L.L. Trygg; Evidence of land use impacts of rapid transit system; Transportation, 6 (1977), pp. 231–248</li>
<li><span id='bib27'></span>
[[#bib27|Li and Yeh, 2004]] X. Li, A.G.O. Yeh; Analyzing spatial restructuring of land use patterns in a fast growing region using remote sensing and GIS; Landsc. Urban Plan., 69 (4) (2004), pp. 335–354</li>
<li><span id='bib28'></span>
[[#bib28|Lin and Evans, 2000]] T.C. Lin, A.W. Evans; The relationship between the price of land and size of plot when plots are small[J]; Land Econ., 76 (3) (2000), pp. 386–394</li>
<li><span id='bib29'></span>
[[#bib29|Luca, 1996]] B. Luca; Nodes and places:complexities of railway station redevelopment; Eur. Plan. Stud., 4 (3) (1996), pp. 331–345</li>
<li><span id='bib30'></span>
[[#bib30|Luca and Tejo, 2001]] B. Luca, S. Tejo; Cities on rails: the redevelopment of railway station areas; Routledge, New York (2001)</li>
<li><span id='bib31'></span>
[[#bib31|Makrí and Folkesson, 1999]] Makrí M., Folkesson C. Accessibility Measures for Analyses of Land Use and Traveling with Geographi-cal Information Systems, in: Proceedings of 2nd KFB-Research Conference, Institute of Technology, Lund, 1999.</li>
<li><span id='bib32'></span>
[[#bib32|Morancho, 2003]] A.B. Morancho; A hedonic valuation of urban green areas; Landsc. Urban Plan., 66 (2003), pp. 35–41</li>
<li><span id='bib33'></span>
[[#bib33|Naceur, 2013]] F. Naceur; Impact of urban upgrading on perceptions of safety in informal settlements: case study of Bouakal, Batna; Front. Archit. Res., 2 (2013), pp. 400–408</li>
<li><span id='bib34'></span>
[[#bib34|Paez et al., 2001]] A. Paez, T. Uchida, K. Miyamoto; Spatial association and heterogeneity issues in land price models; Urban Stud., 38 (9) (2001), pp. 1493–1508</li>
<li><span id='bib35'></span>
[[#bib35|Ping and Chen, 2004]] X.Q. Ping, M.Y. Chen; Real estate financing, the price of land and the trend of housing price; World Econ (2004)</li>
<li><span id='bib36'></span>
[[#bib36|Porta et al., 2005]] S. Porta, P. Crucitti, V. Latora; The network analysis of urban streets: a primal approach; Environ. Plan.: Plan. Des., 33 (5) (2005), pp. 705–725</li>
<li><span id='bib37'></span>
[[#bib37|Ratti, 2004]] C. Ratti; Environment and planning B; Plan. Des., 31 (2004), pp. 487–499</li>
<li><span id='bib38'></span>
[[#bib38|Rosenthal and Helsley, 1994]] S.S. Rosenthal, R.W. Helsley; Redevelopment and the urban land price gradient; J. Urban Econ., 35 (2) (1994), pp. 182–200</li>
<li><span id='bib39'></span>
[[#bib39|Ryan, 2011]] S. Ryan; Property values and transportation facilities: finding the transportation-land use connection; J. Plan. Lit., 13 (4) (2011), pp. 412–427</li>
<li><span id='bib40'></span>
[[#bib40|Song and Gao, 2007]] B. Song, B. Gao; A causality test for housing and land prices: 1998–2006; Mod. Econ. Sci (2007)</li>
<li><span id='bib41'></span>
[[#bib41|Su et al., 2008]] J.G. Su, M. Brauer, B. Ainslie, D. Steyn, T. Larson, M. Buzzelli; An innovative land use regression model incorporating meteorology for exposure analysis; Sci. Total Environ., 390 (2) (2008), pp. 520–529</li>
<li><span id='bib42'></span>
[[#bib42|Timm et al., 2004]] H. Timm, C. Borgelt, C. Döring, R. Kruse; An extension to possibilistic fuzzy cluster analysis; Fuzzy Sets Syst., 147 (1) (2004), pp. 3–16</li>
<li><span id='bib43'></span>
[[#bib43|Wang and Yu, 2012]] M. Wang, B. Yu; Landscape characteristic aesthetic structure: construction of urban landscape characteristic time-spatial pattern based on aesthetic subjects; Front. Archit. Res., 1 (2012), pp. 305–315</li>
<li><span id='bib44'></span>
[[#bib44|Ward et al., 1993]] P. Ward, E. Jimenez, G. Jones; Residential land price changes in Mexican cities and the affordability of land for low-income groups; Urban Stud., 30 (9) (1993), pp. 1521–1542</li>
<li><span id='bib45'></span>
[[#bib45|Yang et al., 2012]] J.Y. Yang, Y. Yang, W. Tang; Development of evaluation model for intensive land use in urban centers; Front. Archit. Res., 1 (2012), pp. 405–410</li>
<li><span id='bib46'></span>
[[#bib46|Yano, 2006]] K. Yano; Research by GIS on Changes of Enterprise Location among Stations in Kyoto; Ritsumeikan Literature, Japan (2006)</li>
<li><span id='bib47'></span>
[[#bib47|Yasuhiko, 2000]] G. Yasuhiko; Introduction to Econometrics by Excel; Toyo Economic Press Office, Japan (2000)</li>
<li><span id='bib48'></span>
[[#bib48|Zeng and Zhang, 2006]] X. Zeng, A. Zhang; A study on the relationship between land price and housing price: granger-causality test and its implication; China Land Sci (2006)</li>
<li><span id='bib50'></span>
[[#bib50|Zhu and Wang, 2005]] Zhu Q., Wang J. Optimization research of urban space configuration based on space syntax, Geosci-ence and Remote Sensing Symposium, 2005.</li>
<li><span id='bib49'></span>
[[#bib49|Zhuang and Zhuang, 2007]] Y. Zhuang, X. Zhuang; Impact of subway stations on commercial land prices; China Land Sci., 21 (2007), p. 4</li>
</ol>