CN109508815B - General activity spatial measure analysis method based on subway IC card data - Google Patents

Abstract

The invention proposes a space measurement and analysis method for general school activities based on subway IC card data, which belongs to the field of traffic data mining. This method uses the subway IC card swiping data to measure and analyze the activity space of the general school crowd, and extract the corresponding measurement indicators to obtain a quantitative analysis of the general school group activity space. For the first time, the present invention mines the corresponding general-school travel activity space measurement index from the IC card data, and classifies the general-study population based on the general-study characteristic, and the divided population has significant general-study characteristics. While taking advantage of the objectivity of IC card data, it provides conditions for the research on the formation mechanism of different general learning modes.

Description

Spatial measurement and analysis method of general school activity based on subway IC card data

technical field

The invention relates to a public transportation data mining method, in particular to a method for measuring and analyzing the activity space of subway general students based on subway IC card data.

Background technique

The general school group is an important part of the travel group. In my country, the phenomenon of the separation of residence and school is common among primary and secondary school students, and it is necessary to use means of transportation for medium and long-distance travel. In the travel period of students, the ground traffic is more congested, so the subway has become an important mode of transportation for them. Using the activity space to study the travel behavior of general school groups by means of rail transit can not only understand the temporal and spatial characteristics of their travel behavior, but also facilitate the macro-control of the internal relationship between various activities. In addition, a deeper understanding of the utilization of rail transit and the existing space in the city can be obtained.

Compared with the traditional traffic survey data, the information provided by the subway IC card data has the characteristics of accuracy, large sample size, wide coverage, strong real-time performance and low acquisition cost, which provides a better quality for the study of travelers' spatiotemporal behavior. data base. However, there are many deficiencies in the existing technical research, for example, traditional data are used to measure the personal activity space, and there are defects such as missing travel activity information, inaccurate spatiotemporal information, and short time; The shortcomings of the short data information time, inaccurate characteristic analysis, and the comprehensive characteristics of big data information have not been fully utilized; the existing research mainly explores the activity space of commuter groups, ignoring groups such as primary and secondary school students commuting to school. The location is relatively fixed, and it has a certain impact on the urban spatial structure; a large number of studies at home and abroad combine the activity space with other issues, such as work-living distance, community differentiation and other factors, but pay little attention to the shape of the activity space itself , area, heat and other characteristics and the relationship between travelers and the urban spatial structure, the research results can only analyze specific problems, can not evaluate the urban spatial structure, nor can the traffic generated by the travel behavior of a large proportion of travel groups be analyzed. The combination of load and urban public transportation construction provides a reliable basis for current planning adjustment and future planning. In the prior art, there is no relevant research on the spatial measurement analysis of general school activities.

SUMMARY OF THE INVENTION

Purpose of the invention: Based on the deficiencies of the prior art, the present invention proposes a space measurement and analysis method for general school activities based on subway IC card data.

Technical solution: In order to achieve the above purpose, the method for measuring and analyzing the general school activity space based on the subway IC card data of the present invention utilizes the subway card swiping data to measure and analyze the general school crowd activity space, and extract the corresponding measurement index to analyze the general school group activity space. similarities and differences. The method includes the following steps:

(1) Obtain the subway IC card swiping data and the latitude and longitude data of subway stations, extract the effective information of general school from the card swiping data, match the longitude and latitude data with the geographic map, identify the general school crowd, and establish a general school crowd travel database;

(2) Preprocessing the travel data of general school population;

(3) Analyze the travel time and space characteristics of the general school crowd, and define the activity space of the general school crowd;

(4) Based on the definition of the activity space of the general population, extract the activity space measurement index from the two dimensions of "time" and "space";

(5) Perform cluster analysis on the measurement index data of the general school population activity space, divide the general school population into different categories according to the measurement index of the activity space, and obtain the general school population activity space model based on subway travel.

Preferably, in step (1), the method for matching the spatial location information of subway stations is: using a collector to crawl the latitude and longitude data of subway stations on an electronic map, and using a universal coordinate converter to convert it into data under the WGS-84 coordinate system, Finally, the location information of each subway station is obtained.

Preferably, the trip data preprocessing in step (2) mainly includes the following parts:

21) Extract the travel data of the general school group on working days and holidays (including winter and summer vacations) respectively, and count the travel time, travel start and end points, and travel times of the general school group in different states;

22) Extract the travel data with the school site as the endpoint, match the IC card number, and count the travel time, frequency and endpoint location distribution of the general school group.

Preferably, step (3) includes: comparing the general school crowd information counted in steps (21) and (22) with the corresponding data of other social groups, to obtain the spatiotemporal characteristics of the general school group, and according to the travel distance there is a direction. According to the characteristics of gender differences, the activity space is defined as an ellipse with home as the focus and the length of the major and minor axes calculated according to the distribution of travel points; according to the characteristics of fixed travel time, the frequency of travel and distance between schools are added to define the activity space of the school group.

Preferably, in step (4), the extraction of activity space measure indexes from two dimensions of time and space mainly includes: ① constructing a confidence ellipse, and using flattening and area as the space measure indexes to describe the activity space; ② common-school distance as the space to describe the activity space Measure index; ③ The frequency of general study is used as a time measure index to describe the activity space.

Preferably, in step (4), the spatial coordinates of all active points are used to establish a confidence ellipse and extract the active spatial measurement index, which mainly includes the following steps:

41) Select 95% confidence level to output the confidence ellipse, and obtain the center coordinates, the length of the major and minor semi-axes and the rotation angle of the confidence ellipse formed by the individual activity points;

42) Calculate the confidence ellipse feature index: export the attribute table of the ellipse in the layer, and introduce two indicators of flattening and area to measure and describe the active space. The flattening α of the ellipse can be calculated by the following formula:

Among them, a is the semi-major axis of the ellipse, and b is the semi-minor axis of the ellipse; the larger the value of α, the flatter the ellipse is, the more directional the active point is, and the more likely it is on the same subway line; the smaller the value of α, the more It means that the more round the ellipse is, the more scattered the active points are in space;

The area of the ellipse is calculated by the formula S=π·a·b. The larger the area value, the wider the daily activities of primary and secondary school students;

43) Add the frequency and distance to improve the activity space measurement index, among which,

The frequency of going to school is counted according to the face number of the IC card through the established travel database of the people who go to school;

The distance between schools is calculated using the latitude and longitude information of the site according to the Euclidean distance calculation formula. The formula is as follows:

in,

long1, long2, lat1, and lat2 are the latitude and longitude data of the home and school subway stations, respectively.

Preferably, step (41) specifically includes the following steps:

411) Extract the IC card face number, inbound station number, outbound station number and their corresponding latitude and longitude in the travel database, merge the inbound station number with the outbound station number, and obtain the activity point information of each traveler ;

412) After adding the latitude and longitude data of subway stations in arcgis, convert all coordinates including home and school locations from the WGS84 geodetic reference system to Gauss-Krüger coordinates through projection transformation;

413) Select the direction distribution tool in arcgis, input the transformed coordinates, and select 95% confidence to output the confidence ellipse. Finally, the center coordinates and the length of the major and minor semi-axes of the confidence ellipse composed of a week of activity points for each individual can be obtained. and rotation angle.

Preferably, in step (5), the K-means clustering method is used to classify the activity space, the number of clusters is determined by the elbow rule, and the ratio of the sum of squares between groups to the sum of squares of the overall distance is calculated by using R software, and the larger ratio is taken. As the number of clusters, measure and analyze according to the activity space confidence ellipse and travel frequency and other characteristics, and analyze the common points of the same type of general school travel behavior and the differences between different types.

Beneficial effects:

1. The present invention realizes the measurement and analysis of the activity space of the general school group through the subway IC card data with a long data span, and more comprehensively and accurately reflects the connection between the general school group activity space and the urban spatial layout, in order to solve the problem of urban traffic. The load imbalance problem and the optimization of urban layout provide decision-making basis. It is beneficial to understand the behavioral characteristics of the general school group and the demand for traffic and passenger flow at different times, and to better plan the spatial layout of the city to improve the quality of life of the student group.

2. For the first time, the present invention mines the corresponding spatial measurement index of general school travel activities from the IC card data, and classifies the general school population based on the general school characteristics, and the divided groups have significant general school characteristics. While taking advantage of the objectivity of IC card data, it provides conditions for the research on the formation mechanism of different general learning modes.

3. The method of the present invention carries out quantitative analysis on the spatial measurement. The product of the method is the characteristic value of the confidence ellipse. The confidence ellipse represents the student activity space to a certain extent, and its characteristic value represents the measurement of the activity space. The present invention uses the actual activity range. Converted into a mathematical model, a new data processing mode for home and school sites is proposed, and it is also a new method for analyzing the space of general school activities. Because the focus of the data and the method match very well, it has a strong rationality.

4. The present invention utilizes the subway card swiping data by the method of confidence ellipse, and expresses the general activity space, and at the same time makes up for the deficiencies in the time span and accuracy of the basic analysis data in the existing research, and the limitations and activities of the research objects. The characteristics of the space itself focus on insufficient attention.

Description of drawings

Fig. 1 is the flow chart of the method of the present invention;

2 is a schematic diagram of a clustering result of the elbow rule according to an embodiment of the present invention;

FIG. 3 is a general learning activity space diagram of clustering results according to an embodiment of the present invention.

Detailed ways

The technical solutions of the present invention will be further described below with reference to the accompanying drawings.

Referring to Fig. 1, the method for measuring and analyzing the space of general school activity based on subway IC card data proposed by the present invention includes the following steps:

Step (1), obtain the subway IC card data and the latitude and longitude data of the subway station, extract the effective information of general education from the card swiping data, and match the longitude and latitude data with the geographic map to determine the geographical location of the general school population and the school, and complete the general education. To identify the crowd, establish the subway IC cartoon to learn the crowd travel database.

The data used in the embodiment of the present invention is the data of card swiping on working days for three consecutive weeks in Nanjing from November 2 to 20, 2015 of the Nanjing Metro Company. Nanjing Metro has preferential policies for students, so their ticket cards are of a separate type. First filter out all the travel information of the student card by the condition that the ticket card type is 54. According to "a method of general school identification based on subway card swiping data" (application number: 201711043136.0), the valid travel records of the card type of student card are retained, and the valid travel database of student card is obtained. An example of its data structure is shown in Table 1.

Table 1 Metro IC card data structure

Use the collector to crawl the latitude and longitude data of 122 subway stations on the Baidu map. In order to match the spatial position in the same coordinate system as the Nanjing map, the universal coordinate converter is used to convert it to the data in the WGS-84 coordinate system. , and finally get the location information of each subway station, as shown in Table 2.

Table 2 Longitude and latitude data of subway stations

station number x1 y1 1 118.7128906 32.0112915 2 118.7164917 31.9979248 3 118.7277222 31.9901123 4 118.7382813 31.98449707 5 118.7562866 31.99371338

According to "A Recognition Method for General School Based on Subway Card Swiping Data" (application number: 201711043136.0), delete the records of departure from the school station before 9:00 and 16:00 from the travel records of all identified home and school station card numbers. The record of departure from the home station after that. In addition, since the subway card swiping records of three consecutive working days were used as samples in the method of identifying homes and schools, the card numbers and their records with less than 3 days of school days were deleted, and finally 317,828 school travel records were obtained. So far, the general school population identification has been completed.

Step (2): Preprocess the travel data of the general school population, which is convenient for the analysis of travel behavior characteristics, so as to extract the significant distinguishing points between the general school population and other groups, and improve the definition of the general school population activity space. The data preprocessing mainly includes:

21) Extract the travel data of the general school group on working days and holidays (including winter and summer vacations) respectively, and count the travel time, travel start and end points, and travel times of the general school group in different states;

22) Extract the travel data with the school site as the endpoint, match the IC card number, and count the travel time, frequency and endpoint location distribution of the general school group.

Step (3), analyze the travel time and space characteristics of the general school crowd, and define the activity space of the general school crowd.

When analyzing the travel time and space characteristics of the general school crowd, the general school crowd information counted in steps (21) and (22) is compared with the corresponding data of other social groups. Specifically, the analysis process includes:

①Comparing the travel behavior of the general school population on weekdays and on holidays, there are significant differences in the distribution of travel time, travel OD, and travel frequency. The travel time on holidays generally lags behind working days, the travel frequency is not fixed, and the regularity of travel OD is not strong;

②Comparing the travel behavior of the general school population and other groups, the travel range and travel time of the general school population are significantly fixed, the card swiping time at the school site is significantly concentrated, and the duration is fixed, the amount of card swiping data is normally distributed within the duration, and the activity range is small. , the activity boundary and regularity are obvious;

③Comparing the travel behavior of school-bound people and other groups of people with the school site as the endpoint, the travel trajectories of school-bound people are significantly concentrated and fixed, the travel generation time is significantly concentrated, and the duration is fixed, and the amount of travel generation is normally distributed within the duration. The frequency of travel is stable, and the travel patterns on working days and holidays are significantly different.

Compared with other social groups, the general school group has significant spatial and temporal characteristics through analysis:

Time characteristics: fixed travel period, fixed duration of each trip, centralized travel data at school sites, fixed peak time, stable duration, and fixed travel frequency;

Spatial characteristics: The travel OD is fixed, the activity range is small and fixed, the appearance order of the activity points is fixed, and the direction of the activity range is obviously different.

It can be seen that compared with other groups in the society, the distribution of travel origin and destination of the general school population has a very high degree of certainty. Based on the unique travel characteristics of the general school population above, the definition of its activity space is improved, and the common travel characteristics of the general school population and other social groups are not used as the main indicators.

The activity space emphasizes the objective movement and activities of travelers under the constraints of time and space. The main activities of the general school group are carried out based on school and home, and there are few activity locations, which cannot be determined according to the density distribution of travel origin and destination. At the same time, the travel path of the subway travel is in the underground space, and a buffer zone cannot be formed near the high-density starting and ending points, so it is impossible to define the activity space according to the travel path and buffer zone. Due to the concentration of activity points and strong directionality of the general school population, there are directional differences in the home-based activity distance. Therefore, when defining the activity space of the general school population, for such travelers with obvious directional differences in travel distance, the geometric shape of the activity pattern is defined as follows: The home is the focal point, and the length of the major and minor axes is an ellipse calculated from the distribution of travel points. At the same time, because the general school group has strong travel time and space characteristics compared with other groups, its frequency and distance are included in the definition of the general school group activity space, and the geometric shape of the activity pattern together represents the general school group activity space.

(4) Based on the definition of the activity space of the general population, the activity space measurement indicators are extracted from the two dimensions of "time" and "space". The activity space measurement indicators include: (1) constructing confidence ellipses, and using flattening and area as the spatial measurement indicators to describe the activity space; (2) the distance between schools as a spatial measurement indicator to describe the activity space;

Primary and secondary school students generally have a small range of activities, especially on weekends, they may not travel or use the subway to travel, so the confidence ellipse is not constructed according to working days and rest days, but uses all travel points of the individual within a week. In the construction process, only the coordinates of each individual spatial activity point are needed, and there is no need to distinguish between incoming and outgoing stations. The confidence ellipse is output with a confidence level of 95%, and finally, the center coordinates, the length of the major and minor semi-axes, and the rotation angle of the confidence ellipse composed of a circle of activity points for each individual can be obtained. The 95% confidence ellipse is the result that can be directly calculated based on the latitude and longitude data of key activity points. Ellipses with 68%, 95%, and 99% confidence levels can be selected to represent different standard deviation numbers, which can be adapted according to actual research. For performance analysis, in the examples, the number of standard deviations was taken as 2, and the confidence level of the ellipse was 95%.

The flattening α of an ellipse can be calculated by the following formula:

Among them, a is the major semi-axis of the ellipse, and b is the minor semi-axis of the ellipse. The larger the value of α, the more flat the ellipse is, the more directional the active points are, and the more likely they are on the same subway line; the smaller the value of α, the more round the ellipse is, and the more scattered the active points in space.

The area of the ellipse is calculated by the following formula:

S=π·a·b

The larger the area value, the wider the range of daily activities of primary and secondary school students.

The frequency and distance of passing school are added to improve the activity space measurement index. Based on the objective analysis of a large number of travel data, it is found that the general school group is significantly different from other groups in terms of distance and frequency of travel, so these two points are included in the analysis of the travel time and space characteristics of the general school group as a supplementary measure of the eigenvalue of the confidence ellipse. Among them, the index of the frequency of school trips is obtained by counting the number of times based on the IC card face number in the existing travel database for school trips; the distance between schools is calculated by the Euclidean distance calculation method, and the calculation formula is as follows:

|d ₁₂ |=6368.16×arccos(sinX+cosX)

in,

long1, long2, lat1, and lat2 are the latitude and longitude data of the home and school subway stations, respectively.

(5) Perform cluster analysis on the measurement index data of the activity space of the general school population, and divide the general school population into different categories according to the measurement index of the activity space.

The K-means clustering method is used for cluster analysis, and the K-means clustering method needs to pre-determine the number of clusters before classifying. At the same time, it should be noted that this clustering method mainly calculates the distance between points according to the index value. Therefore, in order to eliminate the influence of the difference in scale of different indicators on the results, it is necessary to standardize the data.

Among the methods for determining the number of clusters, the elbow rule is the most commonly mentioned at present. This method shows the intra-group sum of squares of different cluster numbers. The intra-group sum of squares represents all the data in a class The sum of squares of the distances from the data point to the center point. The smaller the sum of squares within the group, the higher the similarity of each index of each data in the group. The result is shown in Figure 2. It can be seen from the figure that the sum of squares decreases rapidly from one class to five classes, and then decreases relatively slowly. Therefore, 5 or 6 are selected as the number of clusters in the embodiment.

But at the same time, the elbow rule graph only gives the similarity results within the group, and cannot see the difference between the groups. At this time, the population can be divided into 5 categories and 6 categories, and the R software will automatically calculate the ratio of the sum of squares between groups to the sum of squares of the overall distance. The larger the ratio, the greater the difference between different categories. The results in the R software are shown in Table 6. The ratio of the number of clusters to 6 is slightly larger than the ratio of the number of clusters to 5. Therefore, in the embodiment, 6 categories are finally selected as the number of clusters.

Table 3 The ratio of the sum of squares between groups to the sum of squares of the overall distance for different numbers of clusters

Number of clusters Between-group sum of squares/overall distance sum of squares 5 62.4% 6 65.1%

Through the K-means clustering method, the students who go to subway general school are finally divided into six categories, and the average indicators of the six categories of general students are shown in Table 4.

Table 4 The average value of the indicators of the six types of general education groups

A typical individual is selected from each type of general school group, and ArcGIS is used to mark the location of his home, school, and other activity points in the figure, and draw a confidence ellipse representing his activity space according to the activity point, as shown in Figure 3. (a)-(f) correspond to the activity spaces of typical individuals in the six general study groups, respectively. It can be seen that the activity range of each type of typical travel individuals has significantly different characteristics. These confidence ellipses intuitively represent the activity space of the general school group. The eigenvalues of the ellipse are the measurement indicators of the activity space. Classification and analysis can show which groups have a reasonable range of general school activities and which ones are too large, which can reflect the Whether the distribution of schools is reasonable or not, so as to evaluate the existing urban layout and the cooperation between urban traffic and urban layout.

Claims (4)

1. a general learning activity space measurement analysis method based on subway IC card data, is characterized in that, this method may further comprise the steps: (1) Obtain the subway IC card swiping data and the latitude and longitude data of subway stations, extract the effective information of general school from the card swiping data, match the longitude and latitude data with the geographic map, identify the general school crowd, and establish a general school crowd travel database; (2) Preprocessing the travel data of general school population, including: 21) Respectively extract the travel data of the general school group on working days and holidays including winter and summer vacations, and count the travel time, travel start and end points, and travel times of the general school group in different states; 22) Extract the travel data with the school site as the endpoint, match the IC card number, and count the travel time, frequency and endpoint location distribution of the general school group; (3) Analyze the travel time and space characteristics of the general school crowd, and define the activity space of the general school crowd, including: comparing the general school crowd information counted in steps (21) and (22) with the corresponding data of other social groups to obtain the time and space of the general school group. According to the characteristics of directional differences in travel distance, define its activity space as an ellipse with home as the focus and the length of the major and minor axes calculated according to the distribution of travel points; The distance jointly defines the activity space of the general school group; (4) Based on the definition of the activity space of the general school population, the activity space measurement index is extracted from the two dimensions of time and space, in which the flatness and area of the ellipse are used as the spatial measurement index to describe the activity space, and the general school distance is used as the spatial measurement index to describe the activity space. The frequency of general study is used as a time measurement index to describe the activity space, including: 41) Select 95% confidence level to output the confidence ellipse, and obtain the center coordinates, the length of the major and minor semi-axes and the rotation angle of the confidence ellipse formed by the individual activity points; 42) Calculate the feature index of the confidence ellipse, export the attribute table of the ellipse in the layer, and introduce two indicators of flattening and area to measure and describe the active space, wherein the flattening α of the ellipse is calculated by the following formula:

Among them, a is the major semi-axis of the ellipse, and b is the minor semi-axis of the ellipse; The area of the ellipse is calculated by the formula S=π·a·b; 43) Add the frequency and distance to improve the activity space measurement index, among which, The frequency of going to school is counted according to the face number of the IC card through the established travel database of the general population; The distance between schools is calculated using the latitude and longitude information of the site according to the Euclidean distance calculation formula. The formula is as follows: |d ₁₂ |=6368.16×arccos(sinX+cosX) in,

long1, long2, lat1, and lat2 are the latitude and longitude data of the home and school subway stations respectively; (5) Perform cluster analysis on the measurement index data of the general school population activity space, divide the general school population into different categories according to the measurement index of the activity space, and obtain the general school population activity space model based on subway travel. 2. a kind of general school activity space measurement analysis method based on subway IC card data according to claim 1, is characterized in that, in described step (1), the matching method of subway station space position is: utilize collector in electronic map The latitude and longitude data of the subway station is crawled, and the universal coordinate converter is used to convert it into the data under the WGS-84 coordinate system, and finally the location information of each subway station is obtained. 3. a kind of general school activity space measurement analysis method based on subway IC card data according to claim 1, is characterized in that, described step 41) specifically comprises the following steps: 411) Extract the IC card face number, inbound station number, outbound station number and their corresponding latitude and longitude in the travel database, merge the inbound station number with the outbound station number, and obtain the activity point information of each traveler ; 412) After adding the latitude and longitude data of subway stations in arcgis, convert all coordinates including home and school locations from the WGS84 geodetic reference system to Gauss-Krüger coordinates through projection transformation; 413) Select the direction distribution tool in arcgis, input the transformed coordinates, and select 95% confidence to output the confidence ellipse, and finally obtain the center coordinates, the length of the major and minor semi-axes and the rotation angle of the confidence ellipse formed by each individual activity point. 4. a kind of general school activity space measurement analysis method based on subway IC card data according to claim 1, is characterized in that, utilizes K-means clustering method to classify activity space in described step (5), by elbow The number of clusters is determined by the rule, and the ratio of the sum of squares between groups to the sum of squares of the overall distance is calculated by R software, and the larger ratio is taken as the number of clusters, and the measurement analysis is carried out according to the activity space confidence ellipse and the frequency of trips. The commonalities of behaviors and the differences between different types are analyzed.

Images