Online car-hailing and public transport competition relationship analysis method based on built environment

Abstract

The invention belongs to the technical field of data analysis, and particularly relates to an online car-hailing and public transport competition relationship analysis method based on a built environment, and the method comprises the following steps: obtaining online car-hailing data, and carrying out the preprocessing of the online car-hailing data; extracting element data of the built environment; performing spatial correlation test on the built environment element data, and analyzing spatial distribution characteristics; analyzing the online car-hailing and public traffic competition relationship from two aspects of space coverage and time coverage; quantifying the degree of substitution of the online car-hailing to public transportation; an online car-hailing and public transport competition relationship analysis model based on the built environment data is built; and evaluating the model, and checking the performance of the model. According to the method, the defects in the prior art are overcome, the competing relationship between the online-hailed car and the public traffic is analyzed by utilizing the established environment data and the online-hailed car travel order data, the factors influencing the competing relationship between the online-hailed car and the public traffic are explored, and the space-time heterogeneity of the online-hailed car travel demand under the competing relationship is revealed.

Description

Technical Field

[0001] The present invention belongs to the technical field of data analysis, and specifically relates to a method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on a built environment. Background Art

[0002] With the acceleration of urbanization and the diversification of residents' travel needs, urban transportation systems face a series of challenges. Although traditional public transportation systems such as buses and subways can meet the large-scale travel needs of residents to a certain extent, it is difficult to flexibly adapt to the personalized travel needs of residents. At the same time, application-based online car-hailing services, that is, online car-hailing provided by transportation network companies, have ushered in rapid development in China, connecting passengers and private car drivers through mobile applications, providing residents with a convenient travel option.

[0003] With the rise of online ride-hailing services, they are competing with traditional modes of transportation in meeting urban travel needs, which goes against the original purpose of online ride-hailing as a supplement to traditional public transportation.

[0004] At present, existing studies have confirmed that residents' travel choices are affected by the characteristics of the built environment, such as road design, public transportation accessibility, and population density distribution. However, it is still unclear whether the complex competitive relationship between online ride-hailing and public transportation is affected by the built environment. In exploring the competitive relationship between online ride-hailing and taxis, a large number of empirical studies have confirmed the substitution effect of online ride-hailing on taxi business. However, in terms of the competitive relationship between online ride-hailing and public transportation, there is not enough quantitative research to quantify this relationship. In particular, there is a lack of corresponding research considering the spatial differences in different areas and time periods of the city. In addition, traditional analysis methods often ignore the impact of the diversity of the built environment on travel choices. Summary of the invention

[0005] The purpose of the present invention is to provide an analysis method for the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment, which overcomes the shortcomings of the existing technology, uses built environment data and online ride-hailing travel order data to analyze the competitive and cooperative relationship between online ride-hailing and public transportation, explores the factors affecting the competitive and cooperative relationship between online ride-hailing and public transportation, and reveals the spatiotemporal heterogeneity of online ride-hailing travel demand under the competitive and cooperative relationship.

[0006] To solve the above problems, the technical solutions adopted by the present invention are as follows:

[0007] The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment includes the following steps:

[0008] S1. Obtain online car-hailing data and pre-process the online car-hailing data;

[0009] S2, extracting built environment factor data consisting of five dimensions: density, mix, economic factors, design, and public transportation accessibility;

[0010] S3. Conduct spatial correlation tests on built environment element data and analyze their spatial distribution characteristics;

[0011] S4. Analyze the competitive and cooperative relationship between online ride-hailing and public transportation from the perspectives of spatial coverage and temporal coverage;

[0012] S5. Quantify the extent to which online ride-hailing services replace public transportation;

[0013] S6. Construct a competitive and cooperative relationship analysis model between online ride-hailing and public transportation based on built environment data;

[0014] S7. Evaluate the model and test its performance.

[0015] Furthermore, the online car-hailing data described in S1 includes the online car-hailing number, longitude and latitude of the departure and destination, time, and mileage.

[0016] Furthermore, the preprocessing described in S1 utilizes Python software, and the specific steps are as follows:

[0017] (1) Extract key fields from online ride-hailing data;

[0018] (2) Eliminate order data with missing boarding and alighting times, missing longitude and latitude coordinates, or longitude and latitude coordinates outside the research scope;

[0019] (3) Eliminate abnormal travel data with travel time that is too long or too short, or travel distance that is too long or too short.

[0020] Further, the spatial correlation test in S3 adopts the global Moran's index, and the expression of the global Moran's index is:

[0021]

[0022] Where N c is the total number of spatial unit grids, i is the i-th grid, i = 1, 2, 3, n; Z i and Z j are the values ​​of the independent variable at the i-th and j-th grids respectively; Z is the mean of all grid independent variables; W ij is the spatial weight between the i-th and j-th study unit grids;

[0023] The value of Moran's index is between [-1, 1]. Moran_I>0 indicates a clustering trend in space; Moran_I<0 indicates a discrete trend in space; Moran_I=0 indicates that there is no obvious spatial autocorrelation in space.

[0024] Furthermore, the competitive relationship between online ride-hailing and public transportation is analyzed from the perspectives of spatial coverage and temporal coverage as described in S4. The specific steps include:

[0025] (1) Spatial coverage is to determine whether the pick-up and drop-off locations of online-hailing vehicles are within the bus stop and its 400-meter buffer zone, and to determine whether the online-hailing vehicle trip is classified as a trip with the potential to replace public transportation. The expression is:

[0026] S a ={t∈S|O∈q n (400)∩D∈q n (400)}

[0027] In the formula, S a is the set of all online ride-hailing trips identified as potential alternatives; t is each online ride-hailing trip; S is the set of all online ride-hailing trips; O and D are the pickup and drop-off locations, respectively; q n (400) is the buffer area 400 meters away from the bus stop h;

[0028] (2) Time coverage is obtained by comparing the online car-hailing trip time with the public transportation time. If the online car-hailing trip time is shorter than the public transportation time, the online car-hailing trip is regarded as a supplement to the public transportation. The expression is:

[0029] S b ={t∈S a |T a -T b ≤τ t}

[0030] Where: S b is the set of all ride-hailing trips identified as potential substitutes at the secondary level; T a is the travel time of online ride-hailing trip; T b is the travel time for the same trip using public transportation; τ t is the threshold of travel time difference.

[0031] Furthermore, the quantitative expression of the degree of substitution of online ride-hailing for public transportation described in S5 is:

[0032]

[0033] In the formula, R i is the replacement rate of grid i; is the set of online ride-hailing trips that are identified as alternatives in grid i; S i A complete collection of online ride-hailing services for Gridi.

[0034] Furthermore, the competitive relationship analysis model between online ride-hailing and public transportation in S6 adopts a geographically weighted regression GWR model. The GWR model is based on a global regression model. The regression model is expressed by a least squares model, and its expression is:

[0035]

[0036] In the formula, y is the dependent variable of the i-th grid; l is the l-th independent variable; X il is the value of the independent variable l in the i-th grid; U0 is the intercept term; U l represents the regression coefficient of the independent variable l; ε i is the error term at the i-th grid.

[0037] Furthermore, the GWR model adopts a local regression method based on the global regression model, and its expression is:

[0038]

[0039] In the formula, X in represents the independent variables population density, land use mix, housing price, number of bus stops, and road density obtained in step 3, and n is the nth independent variable; R i is the replacement rate of dependent variable grid i obtained in step 6; (x i ,y i ) is the position coordinate of the centroid of the i-th grid; V n (x i ,y i ) is the local regression coefficient of the independent variable n on the i-th grid; ε i Represents random error.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] This paper examines the competitive and cooperative relationship between online ride-hailing and public transportation from the perspective of the built environment, and divides each online ride-hailing trip into potential substitutes or complements based on spatial coverage and temporal coverage. This paper will be applied to online ride-hailing trips in Nanjing, providing empirical data for the debate on whether online ride-hailing and public transportation are substitutes or complements, and also providing theoretical support for the rational allocation of online ride-hailing resources and the formulation of relevant policies. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Flow chart of the analysis method of the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment.

[0043] Figure 2 This is the study area and grid division map of the present invention.

[0044] Figure 3This is a flow chart for determining the competitive relationship between online ride-hailing and public transportation in the present invention.

[0045] Figure 4 This is a schematic diagram of the distribution of public transportation results crawled by the travel time difference inspection crawler of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] like Figure 1 As shown, the present invention discloses a method for analyzing the competitive relationship between online car-hailing and public transportation based on a built environment, comprising the following steps:

[0048] Step 1: Select the research area

[0049] The research area selected in this paper is the main urban area of ​​Nanjing. The map of Nanjing crawled from OpenStreetMap was converted into a SHP file and imported into ArcGIS. The main urban area (Gulou District, Xuanwu District, Jianye District, Qinhuai District, Qixia District and Yuhuatai District) was cut out and then divided into 400m×400m grids using the "Extract by Mask Tool"; Figure 2 shown.

[0050] Step 2: Online car-hailing data processing

[0051] The present invention uses online car-hailing data in the main urban area of ​​Nanjing, which comes from Didi Chuxing Technology Co., Ltd. Among them, data types such as online car-hailing number, latitude and longitude of departure and destination, time, and mileage are included, and the obtained data is processed using Python software.

[0052] First, the data is preprocessed. The nine key fields required by the present invention are extracted from the original data, namely: "VEHICLE_NO", "DEP_TIME", "DEST_TIME", "DEP_LON", "DEP_LAT", "DEST_LON", "DEST_LAT", "DRIVE_MILE", "DRIVE_TIME", namely: "Order number", "Departure time", "Destination arrival time", "Departure longitude", "Departure latitude", "Destination longitude", "Destination latitude", "Driving mileage" and "Driving time". Secondly, the order data with missing boarding and alighting time, missing longitude and latitude coordinates, and longitude and latitude coordinates outside the research scope (main urban area, selected by administrative division code) are eliminated. Finally, the abnormal travel data with too long or too short travel time (greater than 3 hours or less than 1 minute) and too long or too short travel distance (greater than 300km or less than 100m) are eliminated.

[0053] In order to analyze the competitive relationship with public transportation, considering the operating hours of public transportation, we selected the online car-hailing data with operating hours between 6:00 and 22:30 for subsequent analysis.

[0054] Step 3: Extraction of built environment elements

[0055] Since in step one, the study area is an irregular area, which is not conducive to the calculation of various indicators, the divided grid is selected as the indicator for measuring the built environment elements in the study area. The built environment data is obtained based on WorldPop population data, Open Street Map road network and Amap application programming interface (API). In the present invention, population density is selected to represent the density dimension; road density is selected to represent the design dimension; land use mix is ​​selected to represent diversity; in terms of economic factors, housing prices are selected for representation; in terms of public transportation accessibility, the number of buses and stations is selected for representation. The specific calculation method is as follows:

[0056] (1) Population density

[0057] Based on the population density representing the density in the dimension of built environment factors, the population data was extracted from the WorldPop website. The extracted population raster information was imported into ArcGIS, and the “Summary Statistics” tool was used to calculate the sum of the population in each research raster, and then the population density was calculated.

[0058] (2) Road density

[0059] In terms of design, road density is used to characterize the characteristics of the road network. The calculation formula is as follows:

[0060]

[0061] Where D i is the road density within the research unit grid i; Length ik is the length of road k in the research unit grid i; N a is the number of roads in the research unit grid i; Area i is the area of ​​the research unit grid i.

[0062] (3) Land use mix

[0063] The obtained Nanjing land use SHP file was imported into ArcGIS, and the land use type in each research grid was extracted using the fusion and intersection tools, and the land use area of ​​each type in each grid was obtained by computational geometry. In the present invention, the land use mixing degree of each grid was calculated according to the entropy index method, and the calculation formula is as follows:

[0064]

[0065] Where P im is the proportion of land use type m in grid i; N b is the number of land use types in grid i; M i is the land use diversity index in the i-th grid.

[0066] (4) Housing prices

[0067] By crawling relevant real estate platforms such as Lianjia, we obtained some housing price data in the main urban area of ​​Nanjing, and then used the universal kriging interpolation method to obtain the housing price data of the entire main urban area of ​​Nanjing. Finally, ArcGIS was used to calculate the average housing price of each research grid.

[0068] (5) Number of bus stops

[0069] The web crawler code was written in Python, and the Area of ​​Interest (AOI) data, bus routes and station data were obtained through the application programming interface API provided by the Amap electronic map. The obtained AOI data was filtered and reclassified through the ArcGIS platform, and then the number of bus stations in each research grid was counted using the "Summary Statistics Tool" in ArcGIS.

[0070] Step 4: Spatial correlation test

[0071] The spatial correlation test is used to describe the average degree of correlation between all spatial units and surrounding areas in the entire region to analyze their spatial distribution characteristics. The commonly used statistic is the Global Moran's I, which is calculated as follows:

[0072]

[0073] Where N c is the total number of spatial unit grids, i is the i-th grid, i = 1, 2, 3, n; Z i and Z j are the values ​​of the independent variable at the i-th and j-th grids respectively; is the mean of all raster independent variables; W ij is the spatial weight between the i-th and j-th study unit grids;

[0074] The value of Moran's index is between [-1,1]. Moran I>0 indicates a clustering trend in space; Moran I<0 indicates a discrete trend in space; Moran I=0 indicates that there is no obvious spatial autocorrelation in space.

[0075] The null hypothesis of the Moran's index is that the independent variable is randomly distributed in space and is not affected by the surrounding neighboring values. The Z value is usually used as a significance indicator for the Moran's index statistic to test the null hypothesis. The Z test is described as:

[0076]

[0077] Where, Moran I is the observed value of Moran's index; E(Moran I) and Var(Moran I) represent the expectation and variance of the Moran's index, respectively.

[0078] In the present invention, the significance level is set to P<0.1, that is, when Z>1.65 or Z<-1.65, the Moran's index result rejects the null hypothesis, indicating the existence of spatial autocorrelation and non-random distribution.

[0079] Step 5: Competition and cooperation analysis between online ride-hailing and public transportation

[0080] In the present invention, the time coverage and space coverage methods are used to analyze the competitive relationship between online car-hailing and public transportation. The analysis steps are as shown in the attached figure. Figure 3 shown.

[0081] (1) Phase 1: Identifying spatial coverage

[0082] For the first stage of spatial coverage, the purpose is to determine whether ride-hailing has the potential to replace public transportation. Buffer analysis is used to measure transportation coverage, and the area within the buffer zone is considered to be "covered" by the transportation system. Figure 4 shown.

[0083] The first step is to determine the pick-up and drop-off locations of the online ride-hailing trip, and the second step is to check whether these locations are all within the coverage of public transportation, that is, whether they are located within the 400-meter buffer zone of the bus stop. If the pick-up and drop-off locations are all within the coverage of public transportation, the online ride-hailing trip is classified as a trip with the potential to replace public transportation. The model at this level is established as:

[0084] S a ={t∈S|O∈q n (400)∩D∈q n (400)}

[0085] In the formula, S a is the set of all online ride-hailing trips identified as potential alternatives; t is each online ride-hailing trip; S is the set of all online ride-hailing trips; O and D are the pickup and drop-off locations, respectively; q n (400) is the buffer area 400 meters away from the bus stop h.

[0086] (2) Phase 2: Identifying time coverage

[0087] The dichotomy based on the spatial coverage of the first stage may overestimate the substitution effect. The boarding and alighting trips outside the spatial coverage of public transportation are clearly classified as complementary trips at the first level. However, if the travel time saved by online ride-hailing is significant, the trips within the spatial coverage of public transportation should be considered as complementary to public transportation. Therefore, the travel time difference is selected as an additional criterion for identifying potential substitutions in this level of analysis.

[0088] For a public transit cover (t∈S a ), this aspect estimates the travel time of public transportation. The travel time of public transportation needs to be obtained by using the API key obtained by using the Amap open platform, and then obtained through Python crawler programming.

[0089] Among them, "index" is the online car-hailing order number; "duration" is the public transportation travel time; "distance" is the public transportation travel distance; "transits" is the type of public transportation mode selected. Only when the travel time is shortened, online car-hailing is considered a potential substitute for public transportation. If the difference between public transportation and online car-hailing is less than the predefined threshold, it means that the time consumption of public transportation is not significant. That is, it is not easy to reach (the travel cost is high), and travelers will choose online car-hailing.

[0090] The model of this layer is established as follows:

[0091] S b ={t∈S a |T a -T b ≤τ t}

[0092] Where: S b is the set of all ride-hailing trips identified as potential substitutes at the secondary level; T a is the travel time of online ride-hailing trip; T b is the travel time for the same trip using public transportation; τ t is the threshold of travel time difference.

[0093] Step 6: Quantify the substitution effect between online ride-hailing and public transportation

[0094] The degree of substitution at the aggregate level can be achieved by quantifying the substitution effect at the overall level, which is described as follows:

[0095]

[0096] In the formula, R i is the replacement rate of grid i; is the set of online ride-hailing trips that are identified as alternatives in grid i; S i A complete collection of online ride-hailing services for Gridi.

[0097] Step 7: Construct a model for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment

[0098] The present invention adopts the Geographically Weighted Regression (GWR) model, which is an extension of traditional regression analysis and is suitable for data that presents significant geographical differences in space. The GWR model is based on the global regression model, in which it is assumed that the variables are spatially stable and do not change with the change of spatial position, and the regression results obtained are fixed in the entire study area. The global regression model is mainly expressed by the Ordinary Least Squares (OLS) model, which is described as:

[0099]

[0100] In the formula, y is the dependent variable of the i-th grid; l is the l-th independent variable; X il is the value of the independent variable l in the i-th grid; U0 is the intercept term; U l represents the regression coefficient of the independent variable l; ε i is the error term at the i-th grid.

[0101] Studies have shown that variables vary in geographical location, so it is necessary to use local regression methods to deal with this property of spatial data.

[0102] The GWR model is described as:

[0103]

[0104] In the formula, X in represents the independent variables population density, land use mix, housing price, number of bus stops, and road density obtained in step 3, and n is the nth independent variable; R i is the replacement rate of dependent variable grid i obtained in step 6; (x i ,y i ) is the position coordinate of the centroid of the i-th grid; V n (x i ,y i ) is the local regression coefficient of the independent variable n on the i-th grid; ε i Represents random error.

[0105] Step 8: Model Results Evaluation

[0106] Using Adjusted R 2 , AICc value and AIC value were used to test the model performance and compare it with other regression models. 2 The higher the AICc value and AIC value, the better the model fitting performance. The lower the AICc value and AIC value, the more accurate the model is and the higher the feasibility is.

[0107] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment, characterized by: The following steps are involved: S1. Obtain online car-hailing data and pre-process the online car-hailing data; S2, extracting built environment factor data consisting of five dimensions: density, mix, economic factors, design, and public transportation accessibility; S3. Conduct spatial correlation tests on built environment element data and analyze their spatial distribution characteristics; S4. Analyze the competitive and cooperative relationship between online ride-hailing and public transportation from the perspectives of spatial coverage and temporal coverage; S5. Quantify the extent to which online ride-hailing services replace public transportation; S6. Construct a competitive and cooperative relationship analysis model between online ride-hailing and public transportation based on built environment data; S7. Evaluate the model and test its performance.

2. The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment according to claim 1, characterized in that: The online car-hailing data described in S1 includes the online car-hailing number, the latitude and longitude of the departure and destination, time, and mileage.

3. The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment according to claim 2 is characterized by: The preprocessing described in S1 uses Python software, and the specific steps are as follows: (1) Extract key fields from online ride-hailing data; (2) Eliminate order data with missing boarding and alighting times, missing longitude and latitude coordinates, or longitude and latitude coordinates outside the research scope; (3) Eliminate abnormal travel data with travel time that is too long or too short, or travel distance that is too long or too short.

4. The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment according to claim 1, characterized in that: The spatial correlation test in S3 uses the global Moran's index, and the expression of the global Moran's index is: Where N c is the total number of spatial unit grids, i is the i-th grid, i = 1, 2, 3, n; Z i and Z j are the values ​​of the independent variable at the ith and ,th grids respectively; is the mean of all raster independent variables; W ij is the spatial weight between the i-th and j-th study unit grids; The value of Moran's index is between [-1, 1]. Moran_I>0 indicates a clustering trend in space; Moran_I<0 indicates a discrete trend in space; Moran_I=0 indicates that there is no obvious spatial autocorrelation in space.

5. The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment according to claim 1, characterized in that: S4 analyzes the competitive relationship between online ride-hailing and public transportation from the perspectives of spatial coverage and temporal coverage. The specific steps include: (1) Spatial coverage is to determine whether the pick-up and drop-off locations of online-hailing vehicles are within the bus stop and its 400-meter buffer zone, and to determine whether the online-hailing vehicle trip is classified as a trip with the potential to replace public transportation. The expression is: S a ={t∈S|O∈q n (400)∩D∈q n (400)} In the formula, S a is the set of all online ride-hailing trips identified as potential alternatives; t is each online ride-hailing trip; S is the set of all online ride-hailing trips; O and D are the pickup and drop-off locations, respectively; q n (400) is the buffer area 400 meters away from the bus stop h; (2) Time coverage is obtained by comparing the online car-hailing trip time with the public transportation time. If the online car-hailing trip time is shorter than the public transportation time, the online car-hailing trip is regarded as a supplement to the public transportation. The expression is: S b ={t∈S a |T a -T b ≤τ t } Where: S b is the set of all ride-hailing trips identified as potential substitutes at the secondary level; T a is the travel time of online ride-hailing trip; T b is the travel time for the same trip using public transportation; τ t is the threshold value of the travel time difference.

6. The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment according to claim 1, characterized in that: The quantitative expression of the degree of substitution of online ride-hailing for public transportation as described in S5 is: In the formula, R i is the replacement rate of grid i; S i 2 is the set of online ride-hailing trips that are identified as alternatives in grid i; S i A complete collection of online ride-hailing services for Gridi.

7. The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment according to claim 1, characterized in that: The competitive relationship analysis model between online ride-hailing and public transportation in S6 adopts the geographically weighted regression GWR model. The GWR model is based on the global regression model. The regression model is expressed by the least squares model, and its expression is: In the formula, y is the dependent variable of the i-th grid; l is the l-th independent variable; X il is the value of the independent variable l in the i-th grid; U0 is the intercept term; U l represents the regression coefficient of the independent variable l; ε i is the error term at the i-th grid.

8. The method for analyzing the competitive and cooperative relationship between online ride-hailing and public transportation based on the built environment according to claim 7, characterized in that: The GWR model adopts a local regression method based on the global regression model, and its expression is: Where, X in represents the independent variables population density, land use mix, housing price, number of bus stops, and road density obtained in step 3, and n is the nth independent variable; R i is the replacement rate of dependent variable grid i obtained in step 6; (x i ,y i ) is the position coordinate of the centroid of the i-th grid; V n (x i ,y i ) is the local regression coefficient of the independent variable n on the i-th grid; ε i Represents random error.

Images