CN112734532A - Commuting behavior identification method based on shared electric bicycle borrowing and returning point data - Google Patents
Commuting behavior identification method based on shared electric bicycle borrowing and returning point data Download PDFInfo
- Publication number
- CN112734532A CN112734532A CN202110053404.7A CN202110053404A CN112734532A CN 112734532 A CN112734532 A CN 112734532A CN 202110053404 A CN202110053404 A CN 202110053404A CN 112734532 A CN112734532 A CN 112734532A
- Authority
- CN
- China
- Prior art keywords
- travel
- point
- borrowing
- returning
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 239000011159 matrix material Substances 0.000 claims abstract description 25
- 238000012163 sequencing technique Methods 0.000 claims abstract description 11
- 238000004140 cleaning Methods 0.000 claims abstract description 6
- 238000007781 pre-processing Methods 0.000 claims abstract description 5
- 238000011160 research Methods 0.000 claims description 16
- 230000001174 ascending effect Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000012216 screening Methods 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 230000035772 mutation Effects 0.000 claims description 3
- 230000006399 behavior Effects 0.000 description 13
- 238000004458 analytical method Methods 0.000 description 4
- 238000012546 transfer Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 230000001351 cycling effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- YQGOJNYOYNNSMM-UHFFFAOYSA-N eosin Chemical compound [Na+].OC(=O)C1=CC=CC=C1C1=C2C=C(Br)C(=O)C(Br)=C2OC2=C(Br)C(O)=C(Br)C=C21 YQGOJNYOYNNSMM-UHFFFAOYSA-N 0.000 description 2
- 238000012552 review Methods 0.000 description 2
- 238000003915 air pollution Methods 0.000 description 1
- 238000005056 compaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0633—Lists, e.g. purchase orders, compilation or processing
- G06Q30/0635—Processing of requisition or of purchase orders
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0645—Rental transactions; Leasing transactions
-
- G06Q50/40—
Abstract
The invention discloses a commuting behavior identification method based on shared electric bicycle borrowing and returning point data, which comprises the following steps of: 1) preprocessing original data of a user order; 2) the same vehicle number data is used for sequencing the 'borrowing or returning time points', and the trip data is matched based on the 'vehicle use state'; 3) calculating the row distance and time of the travel data matching result and cleaning; 4) clustering longitude and latitude data of the taxi borrowing and returning points on the working day by using a DBSCAN algorithm, obtaining and calculating a hot spot area and a centroid coordinate thereof, and automatically generating a Thiessen polygon to demarcate a new travel cell; 5) generating a travel OD matrix by combining the matching results of the borrowing and returning vehicle points and the travel cell; 6) calculating a travel OD matrix traffic flow coefficient of each working day, and identifying the travel smaller than an appointed threshold value as commuting travel; 7) and identifying the occupational region. The shared electric bicycle commuting behavior is effectively identified from the user order data, and the occupational distribution is analyzed, so that the commuting use is improved; rate and scheduling provide a theoretical basis.
Description
Technical Field
The invention belongs to the field of mining of shared electric bicycle borrowing and returning point data, and relates to a commuting behavior identification method based on the shared electric bicycle borrowing and returning point data.
Background
The domestic urbanization process is accelerated to promote population increase and the holding capacity of motor vehicles is increased, so that various problems of urban pollution, road congestion, air pollution and the like are caused, the daily work and life of urban residents are seriously influenced, and the urban urbanization process becomes a huge bottleneck restricting urban development. The public transportation type travel becomes a common target of more and more urban planners, residents and governments, but the problem of the last kilometer is increasingly highlighted, the obtained shared electric bicycle is based on the sharing economic concept and used as a terminal travel mode to meet the personalized travel requirements in future, and the characteristics of high efficiency, flexibility, convenience and fashion are favorable for guiding residents to form green and reasonable travel habits and meeting the travel accessibility of people.
However, due to the fact that the technical means of the existing manual static dispatching mode is backward, low in efficiency and insufficient in accuracy, the using requirements of the shared electric bicycle are difficult to meet, the commuting and traveling requirements of residents cannot be met in the peak period of the morning and evening, the problems of no vehicle and borrowing, vehicle concentration and the like occur, and the operation service quality of the shared electric bicycle is also seriously influenced. Through observation of the data of the borrowing and returning points of the shared electric bicycle, the time distribution of the shared electric bicycle used on weekdays by the user is more concentrated than the time distribution of the shared electric bicycle used on weekends, and the state is particularly reflected in the peak period of morning and evening. Therefore, the problem of "no vehicle can be borrowed, vehicles are concentrated" of the commuter is easily caused at the peak of the morning and evening. Meanwhile, the proportion of the shared electric bicycles around the subway station can influence the use of the vehicles, generally speaking, commuters have more regularity in traveling and more stable in frequency and time of using the shared electric bicycles, and therefore, the discussion of transfer and commuting traveling of the shared electric bicycles has practical significance for solving the problems of unbalanced distribution and difficult vehicle utilization of the shared electric bicycles.
Disclosure of Invention
The purpose of the invention is as follows: the invention provides a commuting behavior identification method based on shared electric bicycle borrowing and returning point data.
The technical scheme is as follows: the invention discloses a commuting behavior identification method based on shared electric bicycle borrowing and returning point data, which comprises the following steps of:
(1) preprocessing user order original data provided by a shared electric bicycle company: extracting effective data information from the original data and deleting abnormal order information;
(2) sequencing the same vehicle number data according to the time points of borrowing or returning vehicles, matching travel data based on the change of the vehicle use state, and acquiring a vehicle travel record database;
(3) extracting travel record data in the morning and evening peak time period of the working day from the travel record database, calculating the time difference of borrowing and returning vehicles (namely travel time) of all the trips and the Euclidean distance between borrowing and returning vehicle places (namely travel distance), and cleaning the travel data according to the travel time and the travel distance;
(4) space clustering is carried out on longitude and latitude data of vehicle borrowing and returning points at early peak or late peak of a working day by using a DBSCAN clustering algorithm to obtain a plurality of hot spot areas of the vehicle borrowing and returning, a centroid coordinate of the hot spot areas is calculated by using an ArcGIS tool, a Thiessen polygon is automatically generated by inputting the centroid coordinate data, and a shared electric vehicle traveling cell is defined according to information such as an original road network, a traffic cell, land property and the like;
(5) generating a borrowing and returning OD matrix including the trip amount of the shared electric vehicle between any two trip cells according to the trip cells corresponding to the borrowing and returning places in the trip record data;
(6) calculating a traffic flow coefficient of the electric bicycle trip OD matrix on each working day, and identifying the traffic trip record with the traffic flow coefficient smaller than an agreed threshold value as a commuting trip;
(7) identifying the place where the shared electric vehicle commuting user works through recording the occurrence time (morning peak or evening peak) of the commuting trip;
further, the step (1) of preprocessing the user order raw data includes:
1.1 screening the effective information comprises: sharing the number of the electric bicycle, the time point of the bicycle borrowing or returning, the longitude of the bicycle borrowing or returning, the latitude of the bicycle borrowing or returning and the using state of the bicycle;
1.2 deleting abnormal order information comprises: the order information with the residual and missing items, the order information with the logic errors, the order information with the longitude and latitude coordinates far away from the research range, and the order information in the non-research time area of the car borrowing or returning time point;
sequencing the same vehicle number data according to a "borrowing or returning time point", matching travel data based on the change of a "vehicle use state", acquiring a vehicle travel record database, wherein the vehicle use state is 0, which represents that the use state is changed from stop to motion, and is recorded as 0-borrowing, and the opposite vehicle use state is 1, which represents that the use state is changed from motion to stop, and is recorded as 1-returning;
the concrete matching steps are as follows: the method comprises the steps of deleting data with vehicle numbers appearing only once in the first step, sorting data with the same vehicle numbers in the second step in an ascending order according to time, judging whether the vehicle states alternately appear as 0 and 1 in the third step, removing rows with the beginning of 1 and the end of 0 independently (not forming a complete row pair), regarding data pairs with the vehicle states of 0-1 in adjacent order data sorted according to the time sequence as complete row record data, and cycling the steps to traverse all data to complete row record data matching.
The research object in the step (3) is commuting behavior recognition, and it is known that commuting trips often occur in working day morning and evening peak periods, so that trip record data in working day morning and evening peak periods (the morning peak period is 7:00-9: 00; the evening peak period is 17:00-19:00) are extracted from the trip record database, the returning time differences (i.e. trip times) of all trips and the Euclidean distances (i.e. trip distances) between returning vehicle borrowing and returning vehicle places are calculated, and the trip data are cleaned according to the trip times and the trip distances;
in the step (3), the euclidean distance (i.e., the real travel distance) between two points of the car borrowing and returning place in one complete travel record is calculated, and the calculation formula is as follows:
wherein D is the distance between two points in the two-dimensional space of a complete trip borrowing and returning vehicle point, r is the radius of the earth, phi1、λ2Latitude, longitude, phi of a complete trip vehicle borrowing point respectively2、λ1Respectively is latitude and longitude of a complete trip returning point;
calculating the time difference (travel time) of the vehicle borrowing and returning, namely the difference between the time of the vehicle returning point and the time of the vehicle borrowing point, wherein the calculation formula is as follows:
Δt＝treturning vehicle-tBorrowing vehicle
Wherein, Δ t is the time difference (i.e. travel time) of borrowing and returning the car, tReturning vehicleTime of return of vehicle, tBorrowing vehicleBorrowing point time for the vehicle.
Cleaning travel data: and screening unreasonable data, namely data with the travel distance of less than 100m or more than 10km and data with the travel time of less than 30s or more than 2 hours from the calculated travel distance and travel time results according to the common travel rule of the shared electric bicycle.
In the step (4), a DBSCAN clustering algorithm is used for carrying out spatial clustering on longitude and latitude data sets of the vehicle borrowing and returning points in the morning and evening of the working day, and the specific steps are as follows:
4.1) DBSCAN Cluster input parameter determination
The DBSCAN clustering algorithm divides an area into clusters according to density, input parameters are a minimum density threshold value (namely minimum point quantity) MinPts in the clusters and a neighborhood radius (namely a clustering radius) epsilon, and the solving steps are as follows:
calculating Euclidean distances of all borrowing and returning points in a research space range, counting and screening the minimum Euclidean distances of each borrowing point, returning point and other borrowing points and returning points, sorting the extracted minimum Euclidean distance data sets in an ascending order, recording the minimum Euclidean distances as a K-distance value set, drawing a K-distance value line graph, selecting a minimum Euclidean distance value corresponding to an inflection point of the K-distance value line graph sequence as a neighborhood radius, and recording the minimum Euclidean distance value as an epsilon, wherein the inflection point is a mutation point in the image;
aiming at the obtained neighborhood radius epsilon in the steps, calculating and researching a buffer area in a space range by taking all the borrowing and returning points as central points, constructing the buffer area by the neighborhood radius epsilon, counting the number of all the borrowing and returning points in the buffer area and sequencing the borrowing and returning points in an ascending order, recording the number value set of the borrowing and returning points, drawing a borrowing and returning point number value broken line graph, and recording the sequence inflection point in the borrowing and returning point number value broken line graph as a minimum density threshold value as MinPts;
4.2) operation step of DBSCAN clustering algorithm
The clustering objects are: borrowing and returning the car point longitude and latitude data set D in the morning peak or the evening peak of the working day, and outputting the result as a data set D clustering result and noise data which do not belong to any cluster; the method comprises the following specific steps: randomly selecting an unprocessed point p from the latitude and longitude data set D of the car borrowing and returning point, and if the point p meets the following conditions: if the number of points in the neighborhood radius epsilon is larger than the minimum density threshold MinPts, temporarily setting a point p as a core point; the second step is that: in the longitude and latitude data set D of the car borrowing and returning point, the point q belongs to the range defined by the neighborhood radius epsilon of the point p, the point p is a core point, the point q is called to be the point p and can be reached based on the direct density of MinPts, and if a point chain p exists1，p2，……pnSatisfy q ═ p1，p＝pnWhen p isi+1From piIf the direct density is reachable, the density of the point q from the point p is considered to be reachable, and all point objects which are reachable with the density of the point p in the returning vehicle point longitude and latitude data set D are selected as a cluster according to the previous definition, namely a cluster category; the third step: if the point alpha belongs to D, when the density of the point alpha can reach the point q and the point p at the same time, the point q and the point p are considered to be connected in density, the steps are repeated for unprocessed points through continuous density connection iteration according to the judgment basis of density connection, and a final cluster is generated, and the points which do not belong to direct density connection, density connection and density connection are marked as noise data;
4.3) generating a travel cell: calculating the mass center coordinate of each cluster by utilizing an ArcGIS tool, automatically generating a Thiessen polygon, namely a group of continuous polygons consisting of vertical bisectors connecting line segments of two adjacent points by inputting mass center coordinate data through the ArcGIS tool, and delimiting a travel cell of the new shared electric vehicle
In the step (5), generating a borrow vehicle OD (O-Origin starting point, D-Destination end point) matrix: the method comprises the steps of sorting all travel cells according to rows and columns, and using the travel amount of a shared electric vehicle between any two travel cells (the travel amount is the travel times of the shared electric vehicle from a starting point to an end point) as an element matrix, wherein the rows are the starting point travel cells, and the columns are the end point travel cells.
Matching the geographic coordinate position attribute (Xp, Yp) of any borrowing and returning point P with the geographic range of a travel cell x, wherein x belongs to A, and A is a set of all travel cells, if the following conditions are met:
and according to the steps, a travel OD matrix (a matrix which takes the travel quantity of the shared electric bicycle between any two travel cells as an element and the travel quantity of the shared electric bicycle between any two travel cells as a sequencing by rows and columns) is determined.
7. The method for identifying a commuting behavior based on shared electric bicycle borrowing and returning point data according to claim 1, wherein the commuting behavior definition rule in the step (6) is to calculate a traffic flow coefficient between each travel cell in the morning and evening peak time period of the working day based on travel OD matrix data, define a judgment threshold value of a commuting trip according to a traffic flow coefficient change rule, and identify a trip record with the traffic flow coefficient smaller than an agreed threshold value as the commuting trip;
the traffic flow coefficient calculation mode is as follows:
where I is the number of observation cycle days (I is 1,2,3 … n, and the set of I is I), each day in the observation cycle forms an OD matrix,
the value of the travel quantity from the starting cell x to the terminal cell y on the ith day, AxyIn order to ensure that the traveling quantity between a starting point cell x and a terminal point cell y in an observation period I day of a traveling cell,
the average value of the travel amount of the travel cell from the starting cell x to the terminal cell y in the observation period I day, BxyThe standard deviation of the output between the starting point cell x and the end point cell y in the observation period I day is obtained, so that the standard deviation of the output in the observation period is divided by the average value to obtain the coefficient of variation CxyRecording the traffic flow coefficient, wherein the traffic flow coefficient reflects the variation degree of the data set, the smaller the traffic flow coefficient is, the more stable the variation degree of the data set is, when the traffic flow coefficient between two traffic cells is smaller than a convention threshold value, the existence of a stable traffic flow between the two traffic cells is recognized, and the existence of a commuting trip between a starting point cell x and a terminal point cell y is considered;
the method for determining the contract threshold comprises the following steps: calculating the traffic flow coefficient of each week, and carrying out sectional statistics on the numerical values; the traffic flow coefficient subsection statistics of a plurality of weeks exist in the observation period, and a traffic flow coefficient column statistical chart of each section of the plurality of weeks is formed after the subsection statistics in sequence; and (3) as the traffic flow coefficient is increased in sections, the difference of the columns in the histogram of each week is increased, and the corresponding value with the sudden change is regarded as an appointed threshold value.
Step (7) identifying the positions of the shared electric vehicle commuting users by distinguishing the occurrence time of the commuting travel records in the morning peak or the evening peak, wherein the identification step is as follows:
7.1) calibrating all commuting trip borrowing and returning vehicle point pairs: after the commute trip is identified between the starting point cell x and the terminal cell y in the step (6), calibrating all the vehicle borrowing points in the starting point cell x and all the vehicle returning points in the terminal cell y as commute trip vehicle borrowing and returning point pairs;
7.2) judging the positions of the commuting users:
firstly, extracting early peak trip records from commuting trip borrowing and returning vehicle point pair data calibrated in the first step, wherein the starting point in the trip record data is a residence place, and the destination point in the trip record data is a working place;
secondly, late-peak travel records are extracted, the starting point in the travel record data is a working place, the destination point is a residence place, and according to the law, the positions of the commuter users are identified by borrowing and returning the points to the data for all calibrated commuting travels.
For example: given that a commuting trip exists between a starting point cell x and an end point cell y in an early peak period, a vehicle borrowing point a belongs to the starting point cell x, a vehicle returning point b belongs to the end point cell y, the vehicle borrowing point a-vehicle returning point b is recorded as a commuting trip vehicle borrowing and returning point pair, the vehicle borrowing point a is a residence place, and the vehicle returning point b is a working place.
Has the advantages that: compared with the prior art, the invention has the remarkable effects that:
1. compared with the existing position relation analysis process adopting questionnaire survey and utilizing public bicycle, bus and subway card swiping data, the position relation analysis method has the advantages that the process is simpler, the required data cost is lower, and the shared electric bicycle has the effects of high convenience, high flexibility, strong accessibility and the like;
2. the identification method provided by the invention is based on a wide data volume, utilizes million-data-grade spatio-temporal data of the borrowing and returning points, relies on software such as Python, Arcgis and the like, and is more objective in identification and high in accuracy;
3. the invention also provides powerful data early-stage support for dispatching, configuration and management of the shared electric bicycle and provides a firm foundation for the research and compaction of commuting behaviors.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a flow chart of DBSCAN algorithm;
FIG. 3 is a K-distance curve diagram of a shared electric bicycle borrowing and returning point;
FIG. 4 is a line graph of a descending order of number of vehicles in a shared electric bicycle buffer;
FIG. 5 is a chart of divisions of a shared electric bicycle travel cell in a region of Haeosin, Ningbo;
FIG. 6 is a diagram of a distribution of OD travel traffic flow coefficients within one week;
FIG. 7 is a histogram of traffic flow coefficient difference over two weeks.
Detailed Description
In order to explain the technical scheme disclosed by the invention in detail, the following is further explained by combining the embodiment and the attached drawings of the specification.
In the embodiment, the adopted data is the data of the shared electric bicycle borrowing and returning points in Ningbo city, the shared electric bicycle data has the characteristics of large data volume, wide information volume, space-time information and the like, and relevant data information can be extracted from the data for research and analysis. The user order raw data contains 10 parts: the method comprises the steps of sharing travel codes of the electric bicycles by a single bar, sharing operation companies to which the electric bicycles belong, sharing serial numbers of the electric bicycles, borrowing or returning time points, borrowing or returning longitudes, borrowing or returning latitudes, whether vehicles have use problems or not, vehicle states, order numbers and warehousing time.
1. According to the requirement of the invention, the original data of the user order is preprocessed, and the effective information screening comprises the following steps: sharing the number of the electric bicycle, the time point of the bicycle borrowing or returning, the longitude of the bicycle borrowing or returning, the latitude of the bicycle borrowing or returning and the using state of the bicycle; deleting the abnormal order information comprises: the order information with the incomplete items, the order information with the logical errors, the order information with the longitude and latitude coordinates far away from the research range, and the order information in the non-research time area of the car borrowing or returning time point are shown in table 1:
TABLE 1 sharing electric bicycle borrowing and returning point data effective information structure table
| Shared electric bicycle number | Time point of borrowing or returning | Longitude of borrowing or returning vehicle | Latitude of borrowing or returning vehicle | Vehicle state |
| 2500288247 | 2019/9/2 0:01:52 | 121.603 | 29.909 | 0 |
| 5190450463 | 2019/9/6 08:18:35 | 121.555 | 29.888 | 1 |
| 5920113042 | 2019/9/7 16:43:55 | 121.639 | 29.900 | 0 |
The initial data of the invention is 1484899 pieces of original data of the user order of the shared electric bicycle from Ningbo city of Ningbo of 9 month 2 to 9 month 15 in 2019, and 1048600 pieces of order data are obtained in total after the original data of the user order are preprocessed.
2. And sequencing the same vehicle number data according to the time points of borrowing or returning vehicles, and matching the travel data based on the change of the vehicle use state (0-borrowing, 1-returning vehicles). The method comprises the steps of deleting data with vehicle numbers appearing only once in the first step, sorting data with the same vehicle numbers in the second step in an ascending order according to time, judging whether the vehicle states alternately appear as 0 and 1 in the third step, removing rows with the beginning of 1 and the end of 0 independently (not forming a complete row pair), regarding data pairs with the vehicle states of 0-1 in adjacent order data sorted according to the time sequence as complete row record data, and cycling the steps to traverse all data to complete row record data matching. For example, the extracted vehicle number 2500288247 is shown in Table 2 as part of a data record:
TABLE 2 sharing the matching example table of the data of the borrowing and returning points of the electric bicycle
| Number of borrowing or returning vehicle | Shared electric bicycle number | Latitude and longitude of borrowing or returning vehicle | Time point of borrowing or returning | Vehicle state |
| 1 | 2500288247 | 121.6027,29.9086 | 2019/9/1 0:00:02 | 1 |
| 2 | 2500288247 | 121.6033,29.9090 | 2019/9/2 7:27:32 | 0 |
| 3 | 2500288247 | 121.6027,29.9085 | 2019/9/2 7:28:57 | 1 |
| 4 | 2500288247 | 121.6033,29.9090 | 2019/9/2 7:49:34 | 0 |
| 5 | 2500288247 | 121.6027,29.9086 | 2019/9/2 7:59:24 | 1 |
| 6 | 2500288247 | 121.6026,29.9054 | 2019/9/2 14:52:37 | 1 |
| 7 | 2500288247 | 121.6041,29.9023 | 2019/9/6 7:06:17 | 0 |
Observation table 2 according to the above steps: the data line of the number 2 of the borrowing or returning point and the data line of the number 3 of the borrowing or returning point are considered to be a complete data line of the borrowing or returning point and the returning point, and the data line of the number 4 of the borrowing or returning point and the data line of the number 5 of the borrowing or returning point are also considered to be a complete data line of the borrowing point and the returning point. Correspondingly, the borrowing or returning spot numbers 1, 6 and 7 are considered as error data to be deleted.
Matching is completed by means of Python based on the above flow, and 524288 pieces of new matching data information are generated, some examples are shown in table 2:
TABLE 3 valid information table after sharing electric bicycle borrowing and returning point data matching
3. The study object is commuting behavior recognition, commuting trips are known to occur frequently in working day morning and evening peak periods, therefore, trip record data in working day morning and evening peak periods are extracted from the trip record data matching results, the time difference of borrowing and returning vehicles (namely trip time) of all trips and the Euclidean distance between borrowing and returning vehicle places (namely trip distance) are calculated, unreasonable data are screened out on the calculated trip distance and trip time results according to the general rule of the shared electric bicycle trips, namely data with the trip distance smaller than 100m or larger than 10km and data with the trip time smaller than 30s or larger than 2 hours. And finally, inputting all the borrowing and returning point longitude and latitude data into ArcGIS software to observe and discover that the main distribution area of the shared electric bicycle is a Ningbo city eosin area, the research range is defined as the Ningbo city eosin area, and 92466 user order data are obtained in total.
4. Since the traditional commuting research minimum space analysis unit is a traffic zone (TAZ), administrative divisions (streets, districts, counties and cities and the like) are generally used, and the areas of the administrative divisions are generally distributedAt 10km2In the above, considering that the shared electric bicycle commuting behavior has direct relevance with the service range, traffic travelers often cannot use the vehicles within a large scale range of residential areas or employment areas, and therefore a large traffic small-area division scale causes errors in the shared electric bicycle commuting characteristic determination. On the premise, the traffic zone (TAZ) division needs to be properly reduced under the traditional division scale, and the space distribution density condition of the origin-destination points of the shared electric single vehicles is considered, so that the method has region division objectivity; the data of the transfer trip can be screened out to ensure that the data of the transfer trip is a residence and a working place for the identification result of the subsequent place of employment. Therefore, the invention utilizes a DBSCAN clustering algorithm to perform spatial clustering on longitude and latitude data of the vehicle borrowing and returning points at morning and evening of a working day to obtain a plurality of hot spot areas of the vehicle borrowing and returning, utilizes an ArcGIS tool to calculate the centroid coordinates of the hot spot areas, automatically generates a Thiessen polygon by inputting the centroid coordinate data, and defines a travel cell of a shared electric vehicle by depending on the information of the original road network, traffic cells, land property and the like, and the specific steps are as follows:
4.1) DBSCAN input parameter determination:
calculating Euclidean distances of all borrowing and returning points in a research space range, counting and screening the minimum Euclidean distances between each borrowing and returning point and all other points, sorting the extracted minimum Euclidean distance data sets in an ascending order, recording the minimum Euclidean distance data sets as K-distance value sets, drawing a K-distance value line graph, selecting minimum Euclidean distance data corresponding to inflection points (mutation points in the image) of the K-distance value line graph sequence as a clustering radius, and recording the minimum Euclidean distance data as epsilon. A K-distance curve graph of the shared electric bicycle borrowing and returning points is shown in fig. 3, and the selected clustering radius epsilon is 0.0025;
and aiming at the clustering radius epsilon obtained in the steps, calculating and researching a space range by taking the car borrowing and returning points as central points, constructing a buffer area by the clustering radius epsilon, counting the number of all the car borrowing and returning points in the buffer area, sequencing the number in an ascending order, recording the number value set of the car borrowing and returning points, drawing a number value line graph of the car borrowing and returning points, and recording the sequence inflection points (abrupt change points in the image) in the number value line graph of the car borrowing and returning points as MinPts. Fig. 4 shows the result of counting the number of vehicles used in the shared electric bicycle buffer in descending order, where the minimum density threshold MinPts is selected to be 45.
4.2) Python implements DBSCAN clustering algorithm: inputting a selected clustering radius epsilon of 0.0025 and a minimum density threshold MinPts of 45, realizing DBSCAN clustering by utilizing Python to borrow the vehicle returning point longitude and latitude data set D from the 9.2-9.15 day early peak, obtaining hot spot areas of the vehicle returning points, and finally obtaining 88 types of clusters, wherein the clustering part results are shown in a table 4.
TABLE 4 DBSCAN clustering results
4.3) inputting longitude and latitude data of borrowing and returning points belonging to different clusters into ArcGIS to calculate cluster centroid points, automatically generating a Thiessen polygon through inputting cluster centroid coordinate data and an ArcGIS tool, locally adjusting peripheral research ranges according to an original road network and traffic cell layout, and obtaining 105 travel cells in total, wherein the division results are shown in figure 5.
5. And on the basis of the travel cell and the original map, importing the shared electric bicycle borrowing and returning point data record into ArcGIS software to match the borrowing and returning point with the travel cell, and generating a travel OD matrix. And matching the geographic coordinate position attribute (Xp, Yp) of the point P with the geographic range of a travel cell x, wherein x belongs to A, and A is a set of all travel cells, if the following conditions are met:
proving that the car borrowing and returning point P falls in the travel cell x and belongs to the travel cell x; and repeating the steps, and calibrating the travel cells to which all the car borrowing and returning points belong by using an ArcGIS tool. The calibration results are shown in Table 5.
TABLE 5 calibration results of the residential areas of the shared electric bicycle borrowing and returning points
And determining the OD traffic output corresponding to the borrowing and returning point for each traffic trip according to the steps, determining a trip OD matrix, and iterating according to the thought to obtain a working day early peak OD matrix as shown in a table 4.
TABLE 6 origin-destination OD matrix table
6. The commuting behavior definition rule is a type of transportation travel activities with periodicity, regularity, timeliness and stability generated among the traffic cells corresponding to the OD matrix during the morning and evening peak, the variation degree of the travel occurrence quantity value and the attraction quantity value is in a small range between the travel cells of each day in the observation period and does not exceed an agreed threshold value, the traffic flow with the stable characteristic is in the morning and evening peak period, the traffic flow is considered to be the traffic flow according with the commuting characteristic, the sequential thought is taken as the basis, the traffic flow coefficient is calculated for the OD matrix of the morning and evening peak, and the first week, namely 9.2 days-9.8 days is taken as an example:
from a review of FIG. 6, it can be seen that: the distribution range of the traffic flow coefficient is [0.1-3.3], wherein the value of more than 0.5 accounts for 92.82%, and the histogram of the segmented statistical results of the traffic flow coefficient of the shared electric bicycle in two weeks of 9.2-9.15 days is shown in fig. 7.
From a review of FIG. 7, it can be seen that: the stable traffic flow formed among the travel cells is Cxy<The difference of the sectional statistical change of the traffic flow coefficient of each week is small at 0.5, Cxy>After 0.5, the difference of the column bodies of the traffic flow coefficient segmented statistical histogram is increased between two weeks, so that 0.5 is selected asAnd judging the traffic flow coefficient appointment threshold of the commuting trip.
7. According to the traffic flow coefficient as the judging condition, when Cxy<When the time period just falls in the morning and evening peak period, the stable traffic flow among the xy travel cells is regarded as the commuting traffic flow. The judgment results are shown in table 4:
table 7 statistical results of traffic flow coefficients of travel cells
| Travel cell x | Travel cell y | Coefficient of traffic flow | Whether or not there is a |
| 21 | 27 | 0.116296761 | Is that |
| 73 | 9 | 0.121286439 | Is that |
| 25 | 2 | 0.152633433 | Is that |
| 66 | 17 | 0.979795897 | Whether or not |
| 44 | 19 | 1.274754878 | Whether or not |
| 67 | 7 | 1.58113883 | Whether or not |
| 88 | 2 | 2.449489743 | Whether or not |
Through the steps, all the vehicle borrowing points in the starting point cell x and all the returning points in the terminal cell y are calibrated to be the commuting trip vehicle borrowing and returning point pairs, and the commuting trip statistical results of the electric bicycle shared by the working days of morning, evening and peak from 9 month to 9 month and 15 days in 2019 are obtained and are shown in table 5.
TABLE 8 morning and evening peak commuting and traveling statistics of working day
8. And performing a place-of-employment identification process according to the place-of-employment identification criterion:
1) in the morning peak, the travel cell x to the travel cell y are identified as commuting travel, and the cell x to which the vehicle borrowing point belongs is regarded as a residence and the cell y to which the vehicle returning point belongs is regarded as a working place. In the example, the electric bicycle is shared with the number 2700020523, because the electric bicycle travels in the morning peak period, the travel cell 73 of the travel is taken as the residence, and the travel cell 9 is taken as the residence;
2) and at the time of late peak, identifying the travel cell x to the travel cell y as commuting travel, and regarding the cell x to which the vehicle borrowing point belongs as a working place and the cell y to which the vehicle returning point belongs as a residence place. In the example, the electric bicycle number 3790995549 is shared, and because the trip is performed in the late peak period, the travel cell 39 of the trip is taken as a working place, and the travel cell 70 is taken as a residential place. The occupational region identification result can provide a prerequisite for the influence degree of indexes such as follow-up land property, road network distribution and environmental data on the occupational region change.
Therefore, the method provided by the invention can accurately, effectively, objectively and properly extract the shared electric bicycle commuting behavior information of different types, has strong practicability and popularization value, and provides follow-up ideas and solid foundations for series researches based on shared electric bicycle transfer and commuting behavior and place identification.
Claims (8)
1. A commuting behavior identification method based on shared electric bicycle borrowing and returning point data is characterized by comprising the following steps:
(1) preprocessing user order original data provided by a shared electric bicycle company: extracting effective data information from the original data and deleting abnormal order information;
(2) sequencing the same vehicle number data according to the borrowing or returning time point, matching the travel data based on the vehicle use state change, and acquiring a vehicle travel record database;
(3) extracting travel record data in the morning and evening peak time period of the working day from the travel record database, calculating borrowing and returning time differences of all trips, namely travel time, and Euclidean distances between borrowing and returning places, namely travel distances, and cleaning the travel data according to the travel time and the travel distances;
(4) carrying out spatial clustering on longitude and latitude data of the vehicle borrowing and returning points at the early peak or the late peak of the working day by utilizing a DBSCAN clustering algorithm to obtain a plurality of hot spot areas of the vehicle borrowing and returning, calculating the centroid coordinate of the hot spot areas by utilizing an ArcGIS tool, automatically generating a Thiessen polygon by inputting the centroid coordinate data and delimiting a new shared electric vehicle traveling cell;
(5) generating a borrowing and returning OD matrix including the trip amount of the shared electric vehicle between any two trip cells according to the trip cells corresponding to the borrowing and returning places in the trip record data;
(6) calculating a traffic flow coefficient of the electric bicycle trip OD matrix on each working day, and identifying the traffic trip record with the traffic flow coefficient smaller than an agreed threshold value as a commuting trip;
(7) the shared electric vehicle commuting user place is identified through the commuting travel recording occurrence time, namely the morning peak or the evening peak.
2. The method for identifying commuting behavior based on shared electric bicycle borrowing and returning point data as claimed in claim 1, wherein the step (1) of preprocessing the user order raw data comprises the following steps:
1.1, extracting effective information including the number of the shared electric bicycle, the time point of the bicycle borrowing or returning, the longitude of the bicycle borrowing or returning, the latitude of the bicycle borrowing or returning and the using state of the bicycle;
and 1.2, deleting one or more of order information with residual or missing items, order information with logic errors, order information with longitude and latitude coordinates far away from the research range and order information with a taxi borrowing or returning time point in a non-research time area.
3. The commuting behavior recognition method based on shared electric bicycle borrowing and returning point data according to claim 1, wherein the same vehicle number data in the step (2) are sorted according to borrowing or returning time points, the trip data are matched based on vehicle use state changes, a vehicle trip record database is obtained, the vehicle use state is 0, which represents that the use state is changed from stop to motion and is recorded as 0-borrowing vehicle, the opposite vehicle use state is 1, which represents that the use state is changed from motion to stop and is recorded as 1-returning vehicle;
the concrete matching steps are as follows: the method comprises the steps of firstly deleting data with vehicle numbers appearing only once, secondly sorting data with the same vehicle numbers in an ascending order according to time, thirdly judging whether the vehicle states alternately appear as 0 and 1, respectively eliminating rows with 1 beginning and 0 ending, fourthly regarding data with the vehicle states of 0-1 in adjacent order data after sorting according to the time sequence as complete travel record data, and circularly traversing all data in the steps to complete travel record data matching.
4. The method for identifying commuting behavior based on shared electric bicycle borrowing and returning point data according to claim 1 or 2, wherein in the step (3), the early peak time period is 7:00-9: 00; the late peak time period is 17:00-19:00, and the time difference of borrowing and returning all trips in the time period is calculated, namely the trip time; and Euclidean distance between the car borrowing and returning places, namely travel distance, and cleaning travel data according to travel time and travel distance;
in the step (3), the euclidean distance between two points of the car borrowing and returning place in the complete travel record, namely the travel distance, is calculated, and the calculation formula is as follows:
wherein D is the Euclidean distance between two points of the borrowing and returning place in the complete travel record, namely the travel distance, r is the radius of the earth, phi1、λ2Latitude, longitude, phi of a complete trip vehicle borrowing point respectively2、λ1Respectively is latitude and longitude of a complete trip returning point;
calculating the time difference of borrowing and returning the vehicle, namely the travel time, namely the difference between the time of returning the vehicle and the time of borrowing the vehicle, wherein the calculation formula is as follows:
Δt＝treturning vehicle-tBorrowing vehicle
Wherein, Δ t is the time difference of borrowing and returning the vehicle, namely the travel time, tReturning vehicleTime of return of vehicle, tBorrowing vehicleBorrowing point time for a vehicle;
cleaning travel data: and deleting the data with the travel distance of less than 100m or more than 10km and the data with the travel time of less than 30s or more than 2 hours from the calculated travel distance and travel time results according to the shared electric bicycle.
5. The commuting behavior recognition method based on shared electric bicycle borrowing and returning point data according to claim 4, wherein in the step (4), the DBSCAN clustering algorithm is used for performing spatial clustering on longitude and latitude data sets of the borrowing and returning point on morning and evening at working day, and the specific steps are as follows:
4.1) DBSCAN Cluster input parameter determination
The DBSCAN clustering algorithm divides the area into clusters according to the density, the input parameters are a minimum density threshold MinPts in the clusters and a neighborhood radius epsilon, and the solving steps are as follows: calculating Euclidean distances of all borrowing and returning points in a research space range, counting and screening the minimum Euclidean distances of each borrowing point, returning point and other borrowing points and returning points, sorting the extracted minimum Euclidean distance data sets in an ascending order, recording the minimum Euclidean distances as a K-distance value set, drawing a K-distance value line graph, selecting a minimum Euclidean distance value corresponding to an inflection point of the K-distance value line graph sequence as a neighborhood radius, and recording the minimum Euclidean distance value as an epsilon, wherein the inflection point is a mutation point in the image;
aiming at the obtained neighborhood radius epsilon in the steps, calculating and researching a buffer area in a space range by taking all the borrowing and returning points as central points, constructing the buffer area by the neighborhood radius epsilon, counting the number of all the borrowing and returning points in the buffer area and sequencing the borrowing and returning points in an ascending order, recording the number value set of the borrowing and returning points, drawing a borrowing and returning point number value broken line graph, and recording the sequence inflection point in the borrowing and returning point number value broken line graph as a minimum density threshold value as MinPts;
4.2) operation step of DBSCAN clustering algorithm
The clustering objects are: borrowing and returning the car point longitude and latitude data set D in the morning peak or the evening peak of the working day, and outputting the result as a data set D clustering result and noise data which do not belong to any cluster; the method comprises the following specific steps: randomly selecting an unprocessed point p from the latitude and longitude data set D of the car borrowing and returning point, and if the point p meets the following conditions: if the number of points in the neighborhood radius epsilon is larger than the minimum density threshold MinPts, the temporary point p is taken as a core point; the second step is that: in the longitude and latitude data set D of the borrowing and returning point, the point q belongs to the point p within the range defined by the neighborhood radius epsilon, and the point pIs the core point, and q is the point p reachable based on the MinPts direct density, if there is a point chain p1，p2，……pnSatisfy q ═ p1，p＝pnWhen p isi+1From piIf the direct density is reachable, the density of the point q from the point p is considered to be reachable, and all point objects which are reachable with the density of the point p in the returning vehicle point longitude and latitude data set D are selected as a cluster according to the previous definition, namely a cluster category; the third step: if the point alpha belongs to D, when the density of the point alpha can reach the point q and the point p at the same time, the point q and the point p are considered to be connected in density, the steps are repeated for unprocessed points through continuous density connection iteration according to the judgment basis of density connection, and a final cluster is generated, and the points which do not belong to direct density connection, density connection and density connection are marked as noise data;
4.3) generating a travel cell: and calculating the mass center coordinate of each cluster by using an ArcGIS tool, and automatically generating a Thiessen polygon, namely a group of continuous polygons consisting of vertical bisectors connecting line segments of two adjacent points by inputting mass center coordinate data through the ArcGIS tool, so as to demarcate a new shared electric vehicle traveling cell.
6. The method for identifying commuting behavior based on shared electric bicycle borrowing and returning point data according to claim 1, wherein in the step (5), a borrowing and returning OD matrix is generated: sequencing all travel cells according to rows and columns, and taking the travel quantity of a shared electric vehicle between any two travel cells as an element matrix, wherein the row is a starting point travel cell number, and the column is an end point travel cell number;
matching the geographic coordinate position attribute (Xp, Yp) of any borrowing and returning point P with the geographic range of a travel cell x, wherein x belongs to A, and A is a set of all travel cells, if the following conditions are met:
and judging that the car borrowing and returning point P falls in a travel cell x and belongs to the travel cell x, calibrating travel cells to which all car borrowing and returning points belong by using an ArcGIS tool according to the steps, determining a travel from a starting point O point of the travel to a destination D point of the travel between the travel cell x and the travel cell y by using the travel cell x to which a travel starting point belongs to the travel cell y to which a destination point belongs, and determining a travel OD matrix by using the travel quantity of the shared electric bicycle between any two travel cells as an element according to the steps, wherein the matrix is a matrix which is formed by sequencing all travel cells in rows and columns and using the travel quantity of the shared electric bicycle between any two travel cells as an element, and the travel quantity is the travel times of the shared electric bicycle from the starting point to the destination point.
7. The method for identifying a commuting behavior based on shared electric bicycle borrowing and returning point data according to claim 1, wherein the commuting behavior definition rule in the step (6) is to calculate a traffic flow coefficient between each travel cell in the morning and evening peak time period of the working day based on travel OD matrix data, define a judgment threshold value of a commuting trip according to a traffic flow coefficient change rule, and identify a trip record with the traffic flow coefficient smaller than an agreed threshold value as the commuting trip;
the traffic flow coefficient calculation mode is as follows:
wherein, I is the set of observation period days, I is 1,2,3 … n, the set of I is I, n is the total number of observation days, each day in the observation period forms an OD matrix,
the value of the travel quantity from the starting cell x to the terminal cell y on the ith day, AxyIn order to ensure that the traveling quantity between a starting point cell x and a terminal point cell y in an observation period I day of a traveling cell,
the average value of the travel amount of the travel cell from the starting cell x to the terminal cell y in the observation period I day, BxyIn order to observe the standard deviation of the output between the starting point cell x and the terminal point cell y in the I day of the period, the standard deviation of the output in the observation period is divided by the average value to obtain the coefficient of variation CxyAnd recording the traffic flow coefficient, and when the traffic flow coefficient between the two traffic cells is smaller than an agreed threshold value, considering that a stable traffic flow exists between the two traffic cells, and regarding the existence of a commuting trip between the starting point cell x and the terminal point cell y.
8. The method for identifying commuting behavior based on shared electric bicycle borrowing and returning point data according to claim 7, wherein the step (7) of identifying the shared electric bicycle commuting user place of residence by distinguishing the occurrence time of the commuting travel record at morning peak or evening peak comprises the steps of:
7.1) calibrating all commuting trip borrowing and returning vehicle point pairs: after the commute trip is identified between the starting point cell x and the terminal cell y in the step (6), calibrating all the vehicle borrowing points in the starting point cell x and all the vehicle returning points in the terminal cell y as commute trip vehicle borrowing and returning point pairs;
7.2) judging the positions of the commuting users:
firstly, extracting early peak trip records from commuting trip borrowing and returning vehicle point pair data calibrated in the first step, wherein the starting point in the trip record data is a residence place, and the destination point in the trip record data is a working place;
secondly, late-peak travel records are extracted, the starting point in the travel record data is a working place, the destination point is a residence place, and according to the law, the positions of the commuter users are identified by borrowing and returning the points to the data for all calibrated commuting travels.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110053404.7A CN112734532A (en) | 2021-01-15 | 2021-01-15 | Commuting behavior identification method based on shared electric bicycle borrowing and returning point data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110053404.7A CN112734532A (en) | 2021-01-15 | 2021-01-15 | Commuting behavior identification method based on shared electric bicycle borrowing and returning point data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112734532A true CN112734532A (en) | 2021-04-30 |
Family
ID=75591576
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110053404.7A Pending CN112734532A (en) | 2021-01-15 | 2021-01-15 | Commuting behavior identification method based on shared electric bicycle borrowing and returning point data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112734532A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240265A (en) * | 2021-05-11 | 2021-08-10 | 西北工业大学 | Urban space division method based on multi-mode traffic data |
| CN113283660A (en) * | 2021-06-04 | 2021-08-20 | 北京市交通信息中心 | Internet rental bicycle regional operation associated feature identification method and system |
| CN115440040A (en) * | 2022-09-02 | 2022-12-06 | 重庆大学 | Commuting vehicle identification method based on highway traffic data |
| CN115830901A (en) * | 2023-02-17 | 2023-03-21 | 交通运输部规划研究院 | Method and device for determining truck stopping point |
| CN117612254A (en) * | 2023-11-23 | 2024-02-27 | 广州大学 | Shared bicycle commute behavior based identification method, system, equipment and medium |
-
2021
- 2021-01-15 CN CN202110053404.7A patent/CN112734532A/en active Pending
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113240265A (en) * | 2021-05-11 | 2021-08-10 | 西北工业大学 | Urban space division method based on multi-mode traffic data |
| CN113240265B (en) * | 2021-05-11 | 2023-10-27 | 西北工业大学 | Urban space division method based on multi-mode traffic data |
| CN113283660A (en) * | 2021-06-04 | 2021-08-20 | 北京市交通信息中心 | Internet rental bicycle regional operation associated feature identification method and system |
| CN113283660B (en) * | 2021-06-04 | 2022-07-05 | 北京市智慧交通发展中心(北京市机动车调控管理事务中心) | Internet rental bicycle regional operation associated feature identification method and system |
| CN115440040A (en) * | 2022-09-02 | 2022-12-06 | 重庆大学 | Commuting vehicle identification method based on highway traffic data |
| CN115440040B (en) * | 2022-09-02 | 2023-09-22 | 重庆大学 | Commuter vehicle identification method based on expressway traffic data |
| CN115830901A (en) * | 2023-02-17 | 2023-03-21 | 交通运输部规划研究院 | Method and device for determining truck stopping point |
| CN117612254A (en) * | 2023-11-23 | 2024-02-27 | 广州大学 | Shared bicycle commute behavior based identification method, system, equipment and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112734532A (en) | Commuting behavior identification method based on shared electric bicycle borrowing and returning point data | |
| Chen et al. | Dynamic cluster-based over-demand prediction in bike sharing systems | |
| CN105206048B (en) | A kind of city dweller colony transfer mode based on traffic OD data finds system and method | |
| CN108596202B (en) | Method for calculating personal commuting time based on mobile terminal GPS positioning data | |
| Zhang et al. | A framework for passengers demand prediction and recommendation | |
| CN107305590B (en) | Urban traffic trip characteristic determination method based on mobile phone signaling data | |
| CN108280550B (en) | Visual analysis method for comparing community division of public bicycle stations | |
| CN112270460A (en) | Goods source station identification method for overweight truck based on multi-source data | |
| CN107609107A (en) | A kind of trip co-occurrence phenomenon visual analysis method based on multi-source Urban Data | |
| CN108122131B (en) | Public bicycle card swiping data-based commuting behavior and job and residence identification method | |
| CN104731963A (en) | Grid path recommending method and system based on internet of vehicle | |
| CN110633307A (en) | Urban public bicycle connection subway space-time analysis method | |
| CN112579718A (en) | Urban land function identification method and device and terminal equipment | |
| CN106295868A (en) | Traffic trip data processing method and device | |
| CN108681741B (en) | Subway commuting crowd information fusion method based on IC card and resident survey data | |
| CN112533140B (en) | Shared bicycle distribution condition evaluation method based on index | |
| Yuan et al. | Taxi high-income region recommendation and spatial correlation analysis | |
| CN116703132B (en) | Management method and device for dynamic scheduling of shared vehicles and computer equipment | |
| US20230245007A1 (en) | User segmentation method based on toll data of expressway electronic toll collection | |
| Bao et al. | Spatiotemporal clustering analysis of shared electric vehicles based on trajectory data for sustainable urban governance | |
| Yang et al. | GIS‐Based Analysis of Railway’s Origin/Destination Path‐Selecting Behavior | |
| Linh et al. | Exploring the spatial transferability of FEATHERS–An activity based travel demand model–For Ho Chi Minh city, Vietnam | |
| CN111833229A (en) | Travel behavior space-time analysis method and device based on subway dependency | |
| Ou et al. | Big-data-based analysis on the relationship between taxi travelling patterns and taxi drivers' incomes | |
| CN112885094A (en) | Method for identifying coordinated development areas of multiple types of operating vehicles |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |