JECET; December 2019- February 2020; Sec. C; Vol.9. No.1, 074-079. E-ISSN: 2278–179X [DOI: 10.24214/jecet.C.9.1.07479.]

Journal of Environmental Science, Computer Science and Engineering & Technology An International Peer Review E-3 Journal of Sciences and Technology

Available online at www.jecet.org Section C: Engineering & Technology Research Article

Clustering of rail transit stations in

Xing Zhaomin, Zhao Xinhao, Guo Aixin, Zhi Yingchong and Zhao Jinbao

Shandong University Of Technology, Zibo,Shandong,

Received: 03 December 2019; Revised: 10 December 2019; Accepted: 22 December 2019

Abstract: In recent years, with the rapid development of China's transportation, urban rail transit has become more and more a means of transportation to share a large volume of traffic. At the same time, the connection modes of rail transit are becoming more and more diverse. However, with the rapid development of intelligent transportation, didi chuxing stands out. The combination of the two can not only save travel time, but also improve travel speed. The effective connection between rail transit and didi chuxing is influenced by many potential factors, among which the type of land use as the source and destination will have a huge impact on the travel motivation. This thesis with orbit traffic of Chengdu and drabs travel big data as the research object, through clustering analysis of rail transit site first divided into residential, commercial, entertainment, schools and hospitals, five main land use types and then drops travel big data filtering, get the site around 200 m within the scope of the travel data and carries on the depth of mining, different land use types hop on and off the site in the morning and evening peak. Keywords: didi chuxing; rail transit; clustering analysis; land use type

INTRODUCTION

With the continuous development of China's social economy and the acceleration of urbanization level, people's economic activities become more and more frequent, accompanied by the rise of travel demand. The continuous increase of the traffic volume makes the traffic supply seriously insufficient or even lagging behind. The sharp increase of traffic demand and the relative lack of traffic facilities aggravate the contradiction between the traffic supply and the traffic demand, which seriously restricts the sustainable development of the city.

74 JECET; December 2019- February 2020; Sec. C; Vol.9. No.1, 074-079. . DOI: 10.24214/jecet.C.9.1.07479.

Clustering … Xing Zhaomin et al.

Urban rail transit is an independent rail transit system, which is no longer affected by road conditions on the ground. It has the advantages of large traffic volume, punctuality, environmental protection, high safety and small footprint, and has become one of the first choices for people to travel. Along with our country urbanization accelerates, as of the beginning of 2016, there have been 40 cities of rail transit construction projects approved by state, operating now total mileage of 3000 kilometers, the year has more than 100 passengers choose to rail transportation, urban rail transit has entered a stage of vigorous development, transportation function increasingly highlighted in the position in the public transport system, is an important way to travel in the process of residents travel [1].Despite the rapid development of urban rail transit alleviate the pressure of road traffic, subway stations coverage is low, but the height of the vehicle comfort is poorer, too crowded, and the special structure and space form of rail transit and land resources utilization situation, determine the rail transit in terms of coverage and wire mesh density to independence from other modes of transport development. Therefore, in order to ensure certain operational efficiency of rail transit, strong support and effective connection and coordination of other transportation modes are necessary.

RESEARCH CONTENT

(1) Didi Chuxing Development:Didi chuxing, as the representative of ride-hailing, can effectively match drivers with passengers according to orders, avoiding the phenomenon that drivers spend too much time on empty cars. It can also shorten the distance a passenger has to travel before getting into a vehicle. Didi chuxing can not only improve the utilization of resources, but also shorten people's travel time, bringing convenience to citizens' travel. Didi chuxing can solve the problem of using private cars to some extent. It can not only save families' private car expenses, but also relieve the traffic pressure to some extent, as shown in figure1.

Figure1: Road hierarchy diagram

(2) Rail Traffic Data: This paper conducts further research on the basis of the rail transit in Chengdu. By the end of 2016, the rail transit line in Chengdu has a total length of 108.52km, the average daily passenger flow is about 2.1535 million, the transfer coefficient is 1.45, and the average ride distance

75 JECET; December 2019- February 2020; Sec. C; Vol.9. No.1, 074-079. . DOI: 10.24214/jecet.C.9.1.07479.

Clustering … Xing Zhaomin et al. is 11.51 km. As of 2016, there are four rail transit lines in Chengdu. In November 2016, about 100 stations of the four Chengdu rail transit lines undertook a total traffic volume of 56.0505 million person-times, this main object of study is the . The specific traffic volume is shown in figure2.

Figure2: Line 1

From the perspective of travel chain, travelers are more inclined to choose the combination of convenient access modes at both ends of rail transit when making unified decisions on the access modes at both ends of rail transit. In other words, travelers are more sensitive to the convenience of access modes at the arrival end than at the departure end, so the car access mode is more popular among people. Therefore, the connection between rail transit and didi chuxing has a very important relationship with the land nature of the place of origin and destination.

RESEARCH METHOD

Based on the above analysis of relevant background knowledge, this paper divides the land use types along the rail transit lines in Chengdu, and analyzes the data samples of didi chuxing in Chengdu, China supported by the gaia plan.

(1) Data Preparation:Up to November 2016, line 1 has opened and put into operation a total of 22 stations, excluding the mutual influence between other lines, taking into account its own factors and the surrounding environment, mainly selecting the following clustering variable. such as: site size, distance from center, number of entrances and exits, site spacing, volume fraction, building density, number of feeder buses. (2) Data Standardization: In the data analysis of the above clustering variables, the dimensionality of various variables will have a certain impact on the final processing results, so the above variables need to be standardized. The data standardization method adopted in this paper is z-score standardization method. A standard score, also known as a z-score, is the difference between a score and the mean divided by the standard deviation. Can be expressed as:

76 JECET; December 2019- February 2020; Sec. C; Vol.9. No.1, 074-079. . DOI: 10.24214/jecet.C.9.1.07479.

Clustering … Xing Zhaomin et al.

z = (xij − μj)/σ푗。 Where, z —the variable value after standardization;

xij — the variable value;

µ푗 — the mean value;

σ푗 — the standard deviation. The quantity of z value represents the distance between the original score and the mean of the matrix, calculated in units of standard deviation. When the original score is lower than the mean, Z is negative and vice versa. The result of the original data normalization is shown in table 1.

Table 1: The result of the original data normalization

编号 ZSco01 ZSco02 ZSco03 ZSco04 ZSco05 ZSco06 ZSco07 F1 -0.58787 -0.28428 -0.08536 1.97793 -1.57706 -3.11324 -2.04853 F2 0.78968 -0.51612 2.73157 1.37773 1.24388 0.14495 2.66975 F3 -0.92543 -0.72863 -1.02434 1.67783 0.15446 0.42224 1.36815 F4 2.12292 -1.01843 1.79259 0.17733 -0.49486 0.56089 0.22926 F5 -0.77235 -1.13434 -1.02434 -1.32317 2.10242 0.49156 -0.74694 F6 1.18607 -1.33527 -0.08536 -1.02307 -0.13413 -0.54828 -0.42154 F7 -0.36412 -1.18999 -0.08536 -1.62327 1.30881 0.63021 -0.25884 F8 -0.13856 -1.03775 -1.02434 -0.72297 -1.36062 0.56089 0.55466 F9 0.27145 -0.82523 1.79259 -0.42287 0.05345 0.63021 0.39196 F10 -0.67064 -0.65135 -1.02434 -0.42287 0.94086 0.69953 -0.42154 F11 -0.9725 -0.43884 -0.08536 0.17733 0.16167 0.49156 0.22926 F12 2.10037 -0.207 0.85362 0.77753 0.15446 0.35292 -0.09614 F13 -1.15968 0.02483 -0.08536 1.07763 0.94808 -0.20167 0.71736 F14 -1.26952 0.25667 -0.08536 -0.42287 -0.8556 0.69953 0.22926 F15 -0.4555 0.41123 -1.02434 -1.53324 -0.92774 0.28359 -0.42154 F16 -0.93008 0.5851 -0.08536 0.56746 -0.49486 -2.62798 -0.74694 F17 -0.07593 0.85558 -0.08536 0.53745 -1.57706 -0.54828 0.71736 F18 0.15418 1.02946 -0.08536 -0.36285 0.94086 0.0063 0.22926 F19 -0.09332 1.24197 -0.08536 0.47743 -1.21633 0.21427 -0.25884 F20 -0.39353 1.47381 -0.08536 0.17733 0.06788 -0.06302 0.71736 F21 1.20164 1.66701 -0.08536 -0.72297 0.33483 0.42224 -0.90964 F22 0.98271 1.82156 -1.02434 -0.42287 0.22661 0.49156 -1.72313

(2) K-Means: In this paper, k-means algorithm is adopted for clustering analysis. According to the distance between samples or the similarity (affinity), samples that are more similar and less different are clustered into one category (cluster). Finally, multiple clusters are formed, so that the samples within the same cluster have high similarity and the differences between different clusters are high. The result of clustering is shown in table 2.

77 JECET; December 2019- February 2020; Sec. C; Vol.9. No.1, 074-079. . DOI: 10.24214/jecet.C.9.1.07479.

Clustering … Xing Zhaomin et al.

Table 2: Cluster Analysis Results

4 Stations of Line 1

1 F15, F17, F18, F19, F21, F22 2 F13, F14, F16, F20 3 F1, F3, F10, F11 4 F4, F6, F7, F9 5 F2, F5, F8, F12

(3) Site Classification: Cluster 1 is a general site, mainly including incubation park, century city, tianfu third street, tianfu fifth street, sihe and guangdu. The main feature of this kind of station is that it is far away from the central business district of the city, and its main function is to serve as a general station for passengers to take and drop off. The land around the station is of low development intensity. Cluster 2 is a single travel station, mainly including gaoxin station, city square, jincheng square and huafu avenue. This kind of station is mainly characterized by small scale and relatively single land use type in the surrounding area. However, it is relatively close to the central area of the city and residents have relatively single travel modes. Cluster 3 is a secondary central station, mainly including shengxian lake, north renmin road, nijiaqiao, tongzi forest. This kind of station is characterized by its moderate distance from the central area of the city, a good environment for walking around, and convenient transfer of public transportation. The passenger flow is increased compared with the first two types of stations. Cluster 4 is the city center site, mainly including manjusri courtyard, tianfu square, jinjiang hotel, provincial stadium. This kind of site is mainly characterized by its proximity to the central area of the city, the multiple and convenient connections around the site, the large scale of the site and the high density of surrounding buildings. Cluster 5 is a travel intensive station, mainly including the north railway station, mulma city, huaxi bar, the south railway station.The characteristics of this kind of station mainly include the north railway station and the south railway station, with obvious traffic hub features, diversified travel modes, more connection lines and larger station size, and sufficient space for passengers to transfer.

CONCLUSION

The acceleration of China's motorization process has increased the number of motor vehicles. People's demand for travel should not only stay in the road widening and construction of peripheral roads, but also reasonably control the number of motor vehicles. Based on the number of existing motor vehicles, reasonable use can keep them stable rather than soaring year by year. As an innovative mode of urban transportation, didi chuxing can provide travelers with more convenient and faster travel experience than traditional travel mode, and is also the most attractive alternative to daily travel for citizens. The mass promotion and use of didi chuxing apps on mobile phones and mobile networks can enable people to get a taxi without leaving their homes, thus saving people's time cost. In addition, for families without private cars, didi can also save some money. In addition, Chengdu is one of the central city in western China, with the sustained and rapid economic development of Chengdu in recent years, with the accelerating process of urbanization and rising living standards of urban

78 JECET; December 2019- February 2020; Sec. C; Vol.9. No.1, 074-079. . DOI: 10.24214/jecet.C.9.1.07479.

Clustering … Xing Zhaomin et al. residents, in order to promote the sustainable development of city, Chengdu has established public transport priority development strategy, promote the building of urban rail transit. Didi's private car can supplement the gap in public transportation, and the two have formed a cooperative and complementary relationship. And its flexibility and accessibility is high, can serve outside the bus service coverage area. Clustering analysis was carried out on metro line 1 station, and five classification results were finally obtained. Thus, subsequent studies can be carried out and further studied based on the above clustering results.

REFERENCES

1. China Urban Rail Transit Association. Overview of Urban Rail Transit Lines in China in 2016 [EB/OL].2017.1. 2. GAIA Initiative. https://outreach.didichuxing.com/research/opendata/en/.2017. 3. Chen Huaijie. Study on the Connection between Ground Conventional Bus and Urban Rail Transit [D]. ChangAn University, 2015. 4. Owolf. https://www.jianshu.com/p/4f032dccdcef 5. Shao yingyu. Research on the attraction range and travel mode selection of passenger flow on Harbin metro line 1 [D]. Harbin: northeast forestry university, 2015. 6. T.Pei, W. Wang, H. Zhang, et al. Density-based clustering for data containing two types of points[J]. International Journal of Geographical Information Systems, 2015, 9(2):175-193.

Corresponding author: Xing Zhaomin, Shandong University of Technology, Zibo,China

79 JECET; December 2019- February 2020; Sec. C; Vol.9. No.1, 074-079. . DOI: 10.24214/jecet.C.9.1.07479.