Resident travel characteristic analysis method based on mobile phone signaling data

Abstract

The invention discloses a resident travel characteristic analysis method based on mobile phone signaling data, and relates to the technical field of urban traffic planning and data processing.The method comprises the steps that firstly, the mobile phone signaling data are grouped and sorted according to users and time, a residence place and a working place are recognized, fixed equipment data are cleaned, permanent resident population is determined, and non-permanent resident population is divided; calculating travel times and distances according to track point states, and screening motion point time to analyze travel time distribution; compared with the traditional resident home-entry survey, the method overcomes the defects of small sample size, large capital investment, long statistical period and the like; the characteristics that mobile phone signaling is large in data size and wide in covering crowd are utilized, more representative data are obtained, powerful support is provided for urban traffic planning and management decision making, and optimization of urban traffic layout, reasonable planning of public traffic facilities and promotion of urban sustainable development are facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of urban traffic planning and data processing, and in particular to a method for analyzing resident travel characteristics based on mobile phone signaling data. Background Art

[0002] Traditional household surveys have significant shortcomings in studying residents' travel characteristics. Their small sample size makes it difficult to fully represent the actual travel patterns of urban residents, resulting in significant limitations in research results. Surveys require significant human, material, and financial resources, requiring significant investment. The statistical cycle is long, often taking a long time from survey implementation to final results, resulting in poor data timeliness. Furthermore, insufficient information is collected, and due to the subjective factors and comprehension abilities of the respondents, the collected information may be incomplete and inaccurate, failing to meet the needs of urban transportation planning and traffic management for data on residents' travel characteristics.

[0003] Based on the above problems, the present invention proposes a method for analyzing residents' travel characteristics based on mobile phone signaling data to solve one or more of the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for analyzing residents' travel characteristics based on mobile phone signaling data to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solutions: A method for analyzing resident travel characteristics based on mobile phone signaling data comprises the following steps: S1. Sort the mobile phone signaling data by user and time, filter out signaling data whose residence time exceeds a preset threshold, and after validity confirmation, determine the user's residence based on the data exceeding the preset time period; S2. Referring to the method for determining residence, identify the user's workplace based on working hours; S3. Read user data in chronological order, identify the status of the track points, extract the trip OD based on the time interval of consecutive stop points, record the stop point location, and then read the next user data and repeat the above operation; S4. Cleaning data for fixed equipment with unchanged locations, identifying the permanent population based on the criterion of daily stay exceeding 7 hours and accounting for ≥60% of the days; S5. After the permanent population is identified, the non-resident population is divided into transit and migrant workers according to the length of stay standard; S6. Calculate the number of daily trips for a single user based on the state change of the stay field in the trajectory point recognition result table, and obtain the total number of daily trips and the average number of daily trips for the user; S7. Calculate travel distance using the determined stop locations. First, calculate the average travel distance of a single user in a day, and then calculate the total daily, daily average, and trip average travel distances for all users. S8. Filter the movement point time to form a set, set the time period parameters to filter the database records, calculate the travel distribution in different time periods, and complete the analysis of residents' travel characteristics in the time dimension.

[0006] The method of grouping and sorting the mobile phone signaling data by user and time, screening out the signaling data whose residence time exceeds a preset threshold during the residence period, and determining the user's residence based on the data exceeding the preset time period after validity confirmation includes the following steps: S1.1: Mobile phone signaling data is grouped by user and time, and sorted in chronological order. In this way, the signaling data generated by each user at different times is integrated and categorized. The categorization types are divided into time dimension categorization, behavior status categorization, and geographic location categorization. S1.2: Based on the classification results of the signaling data, select 23:00 to 5:00 the next day as the residence period, set a preset threshold for residence time, and filter out points within this residence period whose residence time exceeds the preset threshold; S1.3: Verify the validity of the filtered results. First, check the filtered base station location information and filter out areas where the base station location conflicts with the actual geographical location and is uninhabitable. Then, filter out abnormal and missing data. Finally, based on the planning layout of the urban functional areas, filter out areas where the functional location is inconsistent with the residential behavior. Then, further verify the validity of the residence point data in this area and exclude data that does not conform to the residential logic. S1.4: Collect signaling data from users for more than 10 working days. On the premise that the base station data has passed the validity confirmation, if the user's residence time at the same base station representing the user's residence during the ten working days is greater than the set threshold, the base station will be identified as the user's residence.

[0007] The step S1.1 includes the following contents: The classification by time dimension is to integrate the signaling data of the same user at different dates and times in chronological order, and classify them according to the residential period of 23:00-5:00 the next day and the working period of 9:00-12:00 and 14:00-17:00, so as to identify the place of residence and the place of work; The classification by behavior status is based on the status of the trajectory points in the signaling data; the number of user trips is calculated by counting the number of changes in the stay value, and this is used to determine the status of the trajectory points when analyzing the trip OD; The classification based on geographical location is to classify the signaling data related to the same base station and the base stations within a radius of 500 meters into one category; the base station that meets the specific residence time condition, the time condition is that the residence time during the residential and working hours is greater than the set threshold, and there is a residence situation that meets the requirements within ten working days, is used to identify the user's residence or workplace; at the same time, the data is classified based on the location of the user's stay point, which is determined by the base station location and is used to calculate the travel distance.

[0008] The method for determining the residence by reference, taking the working hours 9:00-12:00 and 14:00-17:00 as the standard, identifying the user's workplace includes the following steps: S2.1: Based on the classification results of the signaling data, select 9:00-12:00 and 14:00-17:00 as the working hours and set a preset dwell time threshold. During these working hours, filter out points where the dwell time of mobile phone users at the same base station and base stations within a radius of 500 meters exceeds the preset threshold. S2.2: Confirm the validity of the filtered results. First, check the filtered base station location information and filter out the base station location that is inconsistent with the actual geographical environment and is in an area where work activities cannot be carried out; then filter out abnormal and missing data; finally, based on the planning layout of the urban functional area, filter out areas where the functional positioning is inconsistent with work behavior, and then further verify the validity of the residence point data in this area to exclude data that does not conform to the residential logic; S2.3: Collect the user's signaling data for more than 10 working days. On the premise that the base station data has passed the validity confirmation, if the user's residence time at the same base station representing the work location during the ten working days is greater than the set threshold, the base station is identified as the user's work location.

[0009] The process of reading user data in chronological order, determining the status of track points, extracting the trip OD based on the time intervals of consecutive stop points, recording the stop point locations, and then reading the next user data and repeating the above operations includes the following steps: S3.1: Read the current user's mobile phone signaling data in the order of event triggering time; S3.2: Determine the state of the track point. First, set the initial state and determine that each user's initial track point is in the stop state. Then calculate the spatial distance between the track point at the current time t and the previous time t-1. Determine the state of the track point at the current time based on the track point state at the previous time and the comparison result of the current distance with the corresponding distance threshold. The track point status is divided into two states: stay and move. When the previous moment was stay and the distance is greater than the distance threshold, the current moment is determined to be stay and the position is the same as the previous moment; when the distance is greater than the distance threshold, the current moment is determined to be temporarily move; when the previous moment was move and the distance is less than the distance threshold, it is determined that both the previous moment and the current moment are stay and the stay position is different from the previous one; when the distance is greater than the distance threshold, the previous moment is determined to be move and the current moment is temporarily move. S3.3: Propose a trip OD based on the time interval of consecutive stay points. First, calculate the maximum time interval between consecutive static stay points in an area to distinguish different travel behaviors and stay states of users and determine the start and end time and location of the trip. Then, data screening and processing are performed. When the maximum time difference is less than the set time threshold, the stop points in this sequence are judged to be short-term trips and are eliminated. When the maximum time difference is greater than the time threshold, cluster analysis is performed on the consecutive stop points. The start time of the stop point set is determined to be the arrival time of the previous trip, and the end time of the stop point set is determined to be the departure time of the current trip. Finally, the base station with the highest frequency of stop points is used as the starting and ending points of the user's trip. Otherwise, the data group is eliminated and the location information of the stop points is recorded. S3.4: After completing the above operations on the current user's data, read the next user's data and repeat the steps of reading data, determining the status of trajectory points, and extracting travel OD, so as to analyze the data of all target users and generate a trajectory point recognition result table. The trajectory point recognition result table consists of user ID, timestamp, stay field, stay point location and travel OD information, so as to obtain travel OD information of residents in the entire city and specific areas.

[0010] The data of fixed equipment with unchanged cleaning location is used to identify the permanent population based on the criterion that the daily stay time exceeds 7 hours and the proportion of days is ≥ 60%, including the following steps: S4.1: Arrange the mobile phone signaling data in chronological order and continuously observe the information of each data point. Set a period of seven working days. If the location of a device remains unchanged during this period, it is preliminarily determined to be fixed device data. If the location of the device remains unchanged for two periods, it is also determined to be fixed device data and the data is cleaned and removed. S4.2: The identification threshold for permanent residents is a daily stay of more than 7 hours per day, with a proportion of days exceeding 60%. This threshold is based on research on residents' daily activity patterns and analysis of actual data, and is used to distinguish permanent residents from other individuals. S4.3: For the cleaned mobile phone signaling data, count the length of stay of each user on different dates and calculate the proportion of days with a daily stay time of more than 7 hours to the total number of days; if this proportion is ≥ 60%, the corresponding user is identified as a permanent resident; in actual operation, use the traversal and judgment algorithm to traverse all user signaling data, complete the screening and marking of qualified users, and realize the identification of permanent residents.

[0011] After the permanent population is identified, the non-resident population is divided into transit persons and mobile persons according to the length of stay standard, including the following steps: S5.1: After the permanent resident population has been identified, the remaining population will be identified as non-resident population for the next step of classification; S5.2: The non-resident population is classified based on the length of stay. All non-residents whose stay on a particular date is less than three hours are classified as transit persons, and the rest of the non-resident population are classified as migrants. S5.3: For each individual in the non-resident population, check their stay time information on all appearance dates; if the stay time on all appearance dates is less than 3 hours, classify them as transit personnel; otherwise, classify them as mobile personnel; in actual operation, use traversal and judgment algorithms to traverse and compare the stay time data of all non-resident populations to complete the classification.

[0012] Calculating the number of daily trips of a single user based on the state change of the stay field in the trajectory point recognition result table to obtain the total number of daily trips and the average number of daily trips of the user includes the following steps: S6.1: Determine the state of the track point based on the stay field in the track point recognition result table, where stay = 1 indicates stationary and stay = 0 indicates moving; S6.2: For a single user, the process of stay changing from 1 to 0 and then back to 1 represents a completed trip. Therefore, the number of trips per day for a user j can be calculated by counting the number of changes in the stay value. The formula is as follows: ; in, represents the number of trips of user j, n represents the total number of information data of user j, represents the state of the i-th trajectory point of user j, represents the state of the i+1th trajectory point of user j, The absolute value of the difference between the states of two adjacent trajectory points is calculated to determine whether the state of the trajectory point has changed. According to the above formula, the total number of daily trips and the average number of daily trips for all populations can be calculated as follows: ; ; Where m represents the number of valid users.

[0013] The method of calculating the travel distance using the determined stop point locations includes first calculating the average travel distance of a single user in a day, and then calculating the total daily, daily average, and trip average travel distances of all users. The method includes the following steps: S7.1: Based on the obtained user stop locations, calculate the distance between the stop points as the straight-line distance of the trip; for user j, use the distinct statement to obtain the unique value set C of the Lon field. Set C contains the coordinates of all the stop points of user j. Set C can be expressed as C={(Lon i , Lat i )|1≤i≤n}; where n is the number of stay points, Lon represents longitude, and Lat represents latitude; S7.2: Based on the coordinates of the stop points obtained above, calculate the travel distance between user j's stop point i and the stop point i+1 by the formula , the formula is as follows: ; in, Indicates the distance, R is the radius of the earth, which is 6371 kilometers, Lon i and Lat i are the longitude and latitude of the i-th stop point, Lon i+1 and Lat i+1 are the longitude and latitude of the i+1th stop point respectively; S7.3: After obtaining the distance between adjacent stop points, calculate the average travel distance of user j in one day according to the formula , the formula is as follows: ; in, represents the average travel distance of user j. Since there are n-1 distances between n stop points, n-1 represents the number of travel segments in a day; S7.4: Add up the travel distances of all users in a day and calculate the total travel distance according to the formula , the formula is as follows: ; in, represents the total daily travel distance, m is the total number of users, represents the sum of m users, represents the sum of the distances between all adjacent stay points of each user j in one day; S7.5: Then divide the total daily travel distance by the total number of users m and calculate the average daily travel distance according to the formula , the formula is as follows: ; in, is the average daily travel distance, represents the total daily travel distance, m is the total number of users; S7.6: Finally, calculate the average travel distance of all users in a day. First, calculate the total number of trips of all users in a day. Let the number of trips for each user j be F j , then the total number of trips is , and then calculate the average travel distance of all users in a day through the formula , the formula is as follows: ; in, is the average trip distance, represents the total daily travel distance, The total number of trips for all users in a day. This value reflects the average distance of each trip.

[0014] The steps of filtering movement point times to form a set, setting time period parameters to filter database records, calculating travel distribution in different time periods, and completing the analysis of residents' travel characteristics in the time dimension include the following steps: S8.1: Distribute travel time into one-hour time periods, count the number of trips occurring within these time periods, and calculate the travel incidence rate for each time period. The travel incidence rates for each time period are linked together in chronological order to form the time distribution of travel. The travel time distribution objectively reflects the temporal distribution of travel demand. S8.2: First, filter the movement point times to form a set. Use the data table as the operation object and Stay=0 as the query condition. Use the select statement to filter out the Timestamp fields of all movement points, which are recorded as set D. Set D is expressed as D={Timestamp i |1≤i≤x}, where x is the number of all moving points; S8.3: Then set two parameters p and q, satisfying (0 < p < q < 24), to define the time interval; the parameters p and q divide the 24 hours of a day into different time periods; S8.4: Based on the obtained set D, the database records are filtered again with Stay=0 and (p < Timestamp < q) as the conditions to obtain all records that occurred in the time period from p to q, which are recorded as set E. Set E is expressed as E={Timestamp i |1≤i≤z}, where z is the number of elements in set E; S8.5: Finally, calculate the probability H of traveling from time p to time q according to the formula pq , the formula is as follows: ; Among them, 0 < p < q < 24, x is the number of all movement points in a day, and z is the number of movement points generated from the p-th moment to the q-th moment; by calculating the H pq value for different combinations of p and q, the travel occurrence probability at any time period in a day is obtained, and the travel distribution in different time periods is determined accordingly; S8.6: Finally, the travel incidence rates of each time period are arranged in chronological order to form the time distribution of travel, thus completing the analysis of residents' travel characteristics in the time dimension.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. Improve data quality and representativeness: The traditional household survey of residents has a small sample size and is difficult to accurately reflect the travel characteristics of all residents; by using mobile phone signaling data, the present invention can obtain more representative data due to its large user base and coverage of various populations, providing a reliable basis for subsequent planning and management.

[0016] 2. Reduce survey costs and time: The traditional survey faces problems such as large capital investment and long statistical cycles. By using mobile phone signaling data, the present invention reduces the manpower and material resources required for large-scale household surveys, reducing costs; at the same time, the data acquisition and analysis efficiency is higher, and the travel characteristic data of residents can be obtained quickly, such as calculating indicators such as travel times and distances, greatly shortening the statistical cycle compared with the traditional method, making the data more timely and providing support for urban traffic planning and management in a timely manner.

[0017] 3. Accurately identify and deeply analyze travel characteristics: Through a series of steps, the present invention accurately identifies the residential and workplaces of residents, classifies permanent and non-permanent residents, and deeply analyzes travel OD, travel time distribution and other characteristics; in the inter-regional OD analysis, the trajectory point status is accurately judged, and the travel OD information is extracted, which helps to understand the origin and destination points and paths of residents' travel, providing detailed data for optimizing the urban traffic layout and reasonably planning traffic facilities, and improving the scientificity and accuracy of urban traffic management.

[0018] 4. Support urban planning and management decision-making: Comprehensive and accurate residents' travel characteristic data provides key support for urban planning and traffic management; based on the travel distance, frequency and other data obtained from the analysis, the public transportation routes and station layouts can be reasonably planned to improve the utilization efficiency of traffic resources; according to the travel time distribution, targeted traffic diversion strategies can be formulated to relieve traffic congestion. It can also provide a reference for optimizing the layout of urban functional areas and promoting the sustainable development of the city. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1A schematic diagram of a work and residence identification process of a method for analyzing resident travel characteristics based on mobile phone signaling data according to the present invention; Figure 2 A schematic diagram of an OD calculation flow based on mobile phone signaling data of a method for analyzing resident travel characteristics based on mobile phone signaling data of the present invention; Figure 3 This is a schematic diagram of the process of identifying permanent and floating populations in a method for analyzing residents' travel characteristics based on mobile phone signaling data according to the present invention; Figure 4 A schematic diagram of user travel time distribution of a method for analyzing resident travel characteristics based on mobile phone signaling data according to the present invention; Figure 5 The present invention is a method flow chart of a method for analyzing resident travel characteristics based on mobile phone signaling data. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] Example: Figures 1-4 As shown, the present invention provides a technical solution. A method for analyzing resident travel characteristics based on mobile phone signaling data comprises the following steps: S1. Sort the mobile phone signaling data by user and time, filter out the signaling data whose residence time exceeds the preset threshold, and after validity confirmation, determine the user's residence based on the data exceeding the preset time period, such as Figure 1 As shown; S2. Referring to the method for determining the place of residence, the user's work place is identified based on the working hours, such as Figure 1 As shown; S3, read user data in chronological order, determine the status of the trajectory point, extract the trip OD according to the time interval of the consecutive stop points, record the stop point location, and then read the next user data and repeat the above operation, such as Figure 2 As shown; S4. For fixed equipment data with unchanged cleaning locations, identify the permanent population based on the criteria of daily stay time exceeding 7 hours and the proportion of days ≥ 60%, such as Figure 3 As shown; S5. After the permanent population is identified, the non-resident population is divided into transit persons and mobile persons according to the length of stay standard, e.g. Figure 3 As shown; S6. Calculate the number of daily trips for a single user based on the state change of the stay field in the trajectory point recognition result table, and obtain the total number of daily trips and the average number of daily trips for the user; S7. Calculate travel distance using the determined stop locations. First, calculate the average travel distance of a single user in a day, and then calculate the total daily, daily average, and trip average travel distances for all users. S8, filter the movement point time to form a set, set the time period parameters to filter the database records, calculate the travel distribution in different time periods, and complete the analysis of residents' travel characteristics in the time dimension, such as Figure 4 shown.

[0022] The method of grouping and sorting the mobile phone signaling data by user and time, screening out the signaling data whose residence time exceeds a preset threshold during the residence period, and determining the user's residence based on the data exceeding the preset time period after validity confirmation includes the following steps: S1.1: Mobile phone signaling data is grouped by user and time, and sorted in chronological order. In this way, the signaling data generated by each user at different times is integrated and categorized. The categorization types are divided into time dimension categorization, behavior status categorization, and geographic location categorization. S1.2: Based on the classification results of the signaling data, select 23:00 to 5:00 the next day as the residence period, set a preset threshold for residence time, and filter out points within this residence period whose residence time exceeds the preset threshold; S1.3: Verify the validity of the filtered results. First, check the filtered base station location information and filter out areas where the base station location conflicts with the actual geographical location and is uninhabitable. Then, filter out abnormal and missing data. Finally, based on the planning layout of the urban functional areas, filter out areas where the functional location is inconsistent with the residential behavior. Then, further verify the validity of the residence point data in this area and exclude data that does not conform to the residential logic. S1.4: Collect the user's signaling data for more than 10 working days. On the premise that the base station data has passed the validity confirmation, if the user's residence time at the same base station representing the residence is greater than the set threshold during the ten working days, the base station is identified as the user's residence, such as Figure 1 shown.

[0023] The step S1.1 includes the following contents: The classification by time dimension is to integrate the signaling data of the same user at different dates and times in chronological order, and classify them according to the residential period of 23:00-5:00 the next day and the working period of 9:00-12:00 and 14:00-17:00, so as to identify the place of residence and the place of work; The classification by behavior status is based on the status of the trajectory points in the signaling data; the number of user trips is calculated by counting the number of changes in the stay value, and this is used to determine the status of the trajectory points when analyzing the trip OD; The classification based on geographical location is to classify the signaling data related to the same base station and the base stations within a radius of 500 meters into one category; the base station that meets the specific residence time condition, the time condition is that the residence time during the residential and working hours is greater than the set threshold, and there is a residence situation that meets the requirements within ten working days, is used to identify the user's residence or workplace; at the same time, the data is classified based on the location of the user's stay point, which is determined by the base station location and is used to calculate the travel distance.

[0024] The method for determining the residence by reference, taking the working hours 9:00-12:00 and 14:00-17:00 as the standard, identifying the user's workplace includes the following steps: S2.1: Based on the classification results of the signaling data, select 9:00-12:00 and 14:00-17:00 as the working hours and set a preset dwell time threshold. During these working hours, filter out points where the dwell time of mobile phone users at the same base station and base stations within a radius of 500 meters exceeds the preset threshold. S2.2: Confirm the validity of the filtered results. First, check the filtered base station location information and filter out the base station location that is inconsistent with the actual geographical environment and is in an area where work activities cannot be carried out; then filter out abnormal and missing data; finally, based on the planning layout of the urban functional area, filter out areas where the functional positioning is inconsistent with work behavior, and then further verify the validity of the residence point data in this area to exclude data that does not conform to the residential logic; S2.3: Collect the user's signaling data for more than 10 working days. On the premise that the base station data has passed the validity confirmation, if the user's residence time at the same base station representing the work location during the ten working days is greater than the set threshold, the base station is identified as the user's work location, such as Figure 1 shown.

[0025] The process of reading user data in chronological order, determining the status of track points, extracting the trip OD based on the time intervals of consecutive stop points, recording the stop point locations, and then reading the next user data and repeating the above operations includes the following steps: S3.1: Read the current user's mobile phone signaling data in the order of event triggering time; S3.2: Determine the state of the track point. First, set the initial state and determine that each user's initial track point is in the stop state. Then calculate the spatial distance between the track point at the current time t and the previous time t-1. Determine the state of the track point at the current time based on the track point state at the previous time and the comparison result of the current distance with the corresponding distance threshold. The track point status is divided into two states: stay and move. When the previous moment was stay and the distance is greater than the distance threshold, the current moment is determined to be stay and the position is the same as the previous moment; when the distance is greater than the distance threshold, the current moment is determined to be temporarily move; when the previous moment was move and the distance is less than the distance threshold, it is determined that both the previous moment and the current moment are stay and the stay position is different from the previous one; when the distance is greater than the distance threshold, the previous moment is determined to be move and the current moment is temporarily move. S3.3: Propose a trip OD based on the time interval of consecutive stay points. First, calculate the maximum time interval between consecutive static stay points in an area to distinguish different travel behaviors and stay states of users and determine the start and end time and location of the trip. Then, data screening and processing are performed. When the maximum time difference is less than the set time threshold, the stop points in this sequence are judged to be short-term trips and are eliminated. When the maximum time difference is greater than the time threshold, cluster analysis is performed on the consecutive stop points. The start time of the stop point set is determined to be the arrival time of the previous trip, and the end time of the stop point set is determined to be the departure time of the current trip. Finally, the base station with the highest frequency of stop points is used as the starting and ending points of the user's trip. Otherwise, the data group is eliminated and the location information of the stop points is recorded. S3.4: After completing the above operations on the current user data, read the next user data and repeat the steps of reading data, determining the status of trajectory points, and extracting travel OD, so as to analyze the data of all target users and generate a trajectory point recognition result table. The trajectory point recognition result table consists of user ID, timestamp, stay field, stay point location and travel OD information, so as to obtain the travel OD information of residents in the entire city and specific areas, such as Figure 2 shown.

[0026] The data of fixed equipment with unchanged cleaning location is used to identify the permanent population based on the criterion that the daily stay time exceeds 7 hours and the proportion of days is ≥ 60%, including the following steps: S4.1: Arrange the mobile phone signaling data in chronological order and continuously observe the information of each data point. Set a period of seven working days. If the location of a device remains unchanged during this period, it is preliminarily determined to be fixed device data. If the location of the device remains unchanged for two periods, it is also determined to be fixed device data and the data is cleaned and removed. S4.2: The identification threshold for permanent residents is a daily stay of more than 7 hours per day, with a proportion of days exceeding 60%. This threshold is based on research on residents' daily activity patterns and analysis of actual data, and is used to distinguish permanent residents from other individuals. S4.3: For the cleaned mobile phone signaling data, count the length of stay of each user on different days and calculate the proportion of days with a daily stay time of more than 7 hours to the total number of days; if the proportion is ≥ 60%, the corresponding user is identified as a permanent resident; in actual operation, all user signaling data are traversed with the help of traversal and judgment algorithms to complete the screening and marking of qualified users to realize the identification of permanent residents, such as Figure 3 shown.

[0027] After the permanent population is identified, the non-resident population is divided into transit persons and mobile persons according to the length of stay standard, including the following steps: S5.1: After the permanent resident population has been identified, the remaining population will be identified as non-resident population for the next step of classification; S5.2: The non-resident population is classified based on the length of stay. All non-residents whose stay on a particular date is less than three hours are classified as transit persons, and the rest of the non-resident population are classified as migrants. S5.3: For each individual in the non-resident population, check their stay time information on all appearance dates; if the stay time on all appearance dates is less than 3 hours, classify them as transit personnel; otherwise, classify them as mobile personnel; in actual operation, use the traversal and judgment algorithm to traverse and compare the stay time data of all non-resident populations to complete the classification, such as Figure 3 shown.

[0028] Calculating the number of daily trips of a single user based on the state change of the stay field in the trajectory point recognition result table to obtain the total number of daily trips and the average number of daily trips of the user includes the following steps: S6.1: Determine the state of the track point based on the stay field in the track point recognition result table, where stay = 1 indicates stationary and stay = 0 indicates moving; S6.2: For a single user, the process of stay changing from 1 to 0 and then back to 1 represents a completed trip. Therefore, the number of trips per day for a user j can be calculated by counting the number of changes in the stay value. The formula is as follows: ; in, represents the number of trips of user j, n represents the total number of information data of user j, represents the state of the i-th trajectory point of user j, represents the state of the i+1th trajectory point of user j, The absolute value of the difference between the states of two adjacent trajectory points is calculated to determine whether the state of the trajectory point has changed. According to the above formula, the total number of daily trips and the average number of daily trips for all populations can be calculated as follows: ; ; Where m represents the number of valid users.

[0029] The method of calculating the travel distance using the determined stop point locations includes first calculating the average travel distance of a single user in a day, and then calculating the total daily, daily average, and trip average travel distances of all users. The method includes the following steps: S7.1: Based on the obtained user stop locations, calculate the distance between the stop points as the straight-line distance of the trip; for user j, use the distinct statement to obtain the unique value set C of the Lon field. Set C contains the coordinates of all the stop points of user j. Set C can be expressed as C={(Lon i , Lat i )|1≤i≤n}; where n is the number of stay points, Lon represents longitude, and Lat represents latitude; S7.2: Based on the coordinates of the stop points obtained above, calculate the travel distance between user j's stop point i and the stop point i+1 by the formula , the formula is as follows: ; in, Indicates the distance, R is the radius of the earth, which is 6371 kilometers, Lon i and Lat i are the longitude and latitude of the i-th stop point, Lon i+1 and Lat i+1 are the longitude and latitude of the i+1th stop point respectively; S7.3: After obtaining the distance between adjacent stop points, calculate the average travel distance of user j in one day according to the formula , the formula is as follows: ; in, represents the average travel distance of user j. Since there are n-1 distances between n stop points, n-1 represents the number of travel segments in a day; S7.4: Add up the travel distances of all users in a day and calculate the total travel distance according to the formula , the formula is as follows: ; in, represents the total daily travel distance, m is the total number of users, represents the sum of m users, represents the sum of the distances between all adjacent stay points of each user j in one day; S7.5: Then divide the total daily travel distance by the total number of users m and calculate the average daily travel distance according to the formula , the formula is as follows: ; in, is the average daily travel distance, represents the total daily travel distance, m is the total number of users; S7.6: Finally, calculate the average travel distance of all users in a day. First, calculate the total number of trips of all users in a day. Let the number of trips for each user j be F j , then the total number of trips is , and then calculate the average travel distance of all users in a day through the formula , the formula is as follows: ; in, is the average trip distance, represents the total daily travel distance, The total number of trips for all users in a day. This value reflects the average distance of each trip.

[0030] The steps of filtering movement point times to form a set, setting time period parameters to filter database records, calculating travel distribution in different time periods, and completing the analysis of residents' travel characteristics in the time dimension include the following steps: S8.1: Distribute travel time into one-hour time periods, count the number of trips occurring within these time periods, and calculate the travel incidence rate for each time period. The travel incidence rates for each time period are linked together in chronological order to form the time distribution of travel. The travel time distribution objectively reflects the temporal distribution of travel demand. S8.2: First, filter the movement point times to form a set. Use the data table as the operation object and Stay=0 as the query condition. Use the select statement to filter out the Timestamp fields of all movement points, which are recorded as set D. Set D is expressed as D={Timestamp i |1≤i≤x}, where x is the number of all moving points; S8.3: Then set two parameters p and q, satisfying (0 < p < q < 24), to define the time interval; the parameters p and q divide the 24 hours of a day into different time periods; S8.4: Based on the obtained set D, the database records are filtered again with Stay=0 and (p < Timestamp < q) as the conditions to obtain all records that occurred in the time period from p to q, which are recorded as set E. Set E is expressed as E={Timestamp i |1≤i≤z}, where z is the number of elements in set E; S8.5: Finally, calculate the probability H of trips occurring from time p to time q according to the formula pq , and the formula is as follows: ; where 0 < p < q < 24, x is the number of all movement points in a day, and z is the number of movement points generated from time p to time q; by calculating H pq values for different combinations of p and q, obtain the trip occurrence probability for any time period within a day, and thereby determine the trip distribution for different time periods, as Figure 4 shown; S8.6: Finally, arrange the trip occurrence rates for each time period in chronological order to form the time distribution of trips, thereby completing the analysis of residents' trip characteristics in the time dimension. Embodiment

[0031] To analyze the trip characteristics of residents in a certain city, collect mobile signaling data for a month in this city, including user ID, timestamp (Timestamp), base station location (longitude and latitude information, i.e., Lon and Lat fields), and track point status (stay field, 1 indicates staying, 0 indicates moving).

[0032] When determining the place of residence, first use the database to group and sort the mobile signaling data by user ID and timestamp. Then, select 23:00 - 5:00 the next day as the residence time period, set a preset threshold of 3 hours for the staying duration, and filter out the points that stay for more than 3 hours within this time period. Subsequently, conduct validity verification on the filtered results, excluding base station data located in uninhabitable areas such as factories, with abnormal and missing data, and areas where the functional location does not match the residence. Finally, collect the signaling data of each user for more than 10 working days. If a user stays for more than 3 hours during the residence time period of a certain base station in ten working days, identify this base station as their place of residence. User A stays for more than 3 hours during the residence time period of base station X in 12 out of 15 working days, so base station X is the place of residence of user A.

[0033] The method of determining the workplace is similar to that of determining the place of residence. Based on the classification result of the signaling data, select 9:00 - 12:00 and 14:00 - 17:00 as the working time periods, set a preset threshold of 2 hours for the staying duration, filter out the points where the mobile phone users stay for more than 2 hours at the same base station and base stations within a radius of 500 meters, and process the filtered results using the same validity verification method to exclude unreasonable data. Collect the signaling data of more than 10 working days. If a user stays for more than 2 hours during the working time period of a certain base station in ten working days, this base station is their workplace. For example, user B stays for more than 2 hours during the working time period of base station Y in 8 out of 12 working days, so base station Y is the workplace of user B.

[0034] To extract trip ODs, user mobile phone signaling data is read in chronological order. Initial trajectory points are set to a dwelling state. The trajectory point state is determined by calculating the spatial distance between adjacent trajectory points (compared to a distance threshold of 500 meters). For example, if user C was previously dwelling and the current distance is greater than 500 meters, the current state is considered moving. The maximum time interval between consecutive dwell points is then calculated (set to 1 hour). Short trips below this threshold are eliminated. Data above this threshold are clustered to determine the trip's origin and destination points. For example, for user D's set of dwell points, the maximum time interval is 1.5 hours. Clustering determines that the start time is the arrival time of the previous trip, and the end time is the departure time of the current trip. The base station with the highest frequency of occurrence is used as the origin and destination point. After performing these operations on all user data, a trajectory point recognition result table is generated.

[0035] To identify permanent residents, we first sorted mobile phone signaling data chronologically. If a device's location remained unchanged for two seven-day periods, it was identified as a fixed device and removed. We then counted each user's duration of stay on different days and calculated the proportion of days with a stay exceeding seven hours. If user E stayed for more than seven hours on 20 days within a 30-day month (a proportion exceeding 66.7%), we identified him as a permanent resident.

[0036] When categorizing the non-resident population, after determining the permanent population, the remaining users are considered non-residents. Examining the duration of stay for each non-resident individual on all dates, for example, user F stayed for less than three hours on all dates and is therefore classified as a transit user. User G stayed for more than three hours on some dates and is therefore classified as a migrant.

[0037] Calculating the number of trips relies on the stay field in the trajectory point recognition results table. For example, in the data for user H on a particular day, the stay value changes from 1 to 0 and back to 1 five times, resulting in 2.5 trips on that day, rounded up to 3. The total daily trips are calculated by adding the trip counts of all 100 valid users and then dividing by 100 to calculate the average daily trip count.

[0038] To calculate travel distance, taking user I as an example, we first use a database statement to retrieve the unique Lon field values ​​of their stopover points, obtaining their coordinates. We then use a formula to calculate the distances between adjacent stopover points, and then calculate the average travel distance for the day (add the distances of all adjacent stopover points and divide by the number of trip segments). We then calculate the total daily travel distance, average daily travel distance, and average trip distance for all users. Assuming the total daily travel distance for all users is 10,000 kilometers and the total number of users is 200, the average daily travel distance is 50 kilometers. If the total number of trips is 500, the average trip distance is 20 kilometers.

[0039] Analyze the travel time distribution, using one-hour periods as the timeframe, count the number of trips during each period, and calculate the trip rate. Filter the travel time points to form set D, set the parameters p = 7 and q = 8, and filter the database records according to these conditions to obtain set E. Calculate the probability of a trip occurring between 7:00 and 8:00 using the formula. Similarly, calculate the probability for other periods to determine the travel distribution. Finally, arrange the travel rates for each period in chronological order. Analysis reveals travel characteristics related to time, such as a higher rate of travel between 7:00 and 9:00 during the morning rush hour.

[0040] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for analyzing resident travel characteristics based on mobile phone signaling data, characterized by: The following steps are involved: S1. Sort the mobile phone signaling data by user and time, filter out signaling data whose residence time exceeds a preset threshold, and after validity confirmation, determine the user's residence based on the data exceeding the preset time period; S2. Referring to the method for determining residence, identify the user's workplace based on working hours; S3. Read user data in chronological order, identify the status of the track points, extract the trip OD based on the time interval of consecutive stop points, record the stop point location, and then read the next user data and repeat the above operation; S4. Cleaning data for fixed equipment with unchanged locations, identifying the permanent population based on the criterion of daily stay exceeding 7 hours and accounting for ≥60% of the days; S5. After the permanent population is identified, the non-resident population is divided into transit and migrant workers according to the length of stay standard; S6. Calculate the number of daily trips for a single user based on the state change of the stay field in the trajectory point recognition result table, and obtain the total number of daily trips and the average number of daily trips for the user; S7. Calculate travel distance using the determined stop locations. First, calculate the average travel distance of a single user in a day, and then calculate the total daily, daily average, and trip average travel distances for all users. S8. Filter the movement point time to form a set, set the time period parameters to filter the database records, calculate the travel distribution in different time periods, and complete the analysis of residents' travel characteristics in the time dimension.

2. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 1, characterized in that: The method of grouping and sorting the mobile phone signaling data by user and time, screening out the signaling data whose residence time exceeds a preset threshold during the residence period, and determining the user's residence based on the data exceeding the preset time period after validity confirmation includes the following steps: S1.1: Group mobile phone signaling data by user and time, and sort them in chronological order. In this way, the signaling data generated by each user at different times can be integrated and classified; Classification types are divided into time dimension classification, behavior state classification and geographic location related classification; S1.2: Based on the classification results of the signaling data, select 23:00 to 5:00 the next day as the residence period, set a preset threshold for residence time, and filter out points within this residence period whose residence time exceeds the preset threshold; S1.3: Verify the validity of the filtered results. First, check the filtered base station location information and filter out areas where the base station location conflicts with the actual geographical location and is uninhabitable. Then, abnormal and missing data are filtered out; finally, based on the planning layout of urban functional areas, areas where functional positioning is inconsistent with residential behavior are screened, and the validity of the residence point data in the area is further verified to exclude data that does not conform to residential logic; S1.4: Collect signaling data from users for more than 10 working days. On the premise that the base station data has passed the validity confirmation, if the user's residence time at the same base station representing the user's residence during the ten working days is greater than the set threshold, the base station will be identified as the user's residence.

3. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 2, characterized in that: The step S1.1 includes the following contents: The classification by time dimension is to integrate the signaling data of the same user at different dates and times in chronological order, and classify them according to the residential period of 23:00-5:00 the next day and the working period of 9:00-12:00 and 14:00-17:00, so as to identify the place of residence and the place of work; The classification by behavior status is based on the status of the trajectory points in the signaling data; the number of user trips is calculated by counting the number of changes in the stay value, and this is used to determine the status of the trajectory points when analyzing the trip OD; The classification based on geographical location is to classify signaling data related to the same base station and base stations within a radius of 500 meters into one category; Base stations that meet specific residence time conditions, where the residence time during residential and working hours is greater than the set threshold, and there are residences that meet the requirements within ten working days, are used to identify the user's residence or workplace; at the same time, based on the location classification data of the user's stay point, the location is determined by the base station location and is used to calculate the travel distance.

4. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 1, characterized in that: The method for determining the residence by reference, taking the working hours 9:00-12:00 and 14:00-17:00 as the standard, identifying the user's workplace includes the following steps: S2.1: Based on the classification results of the signaling data, select 9:00-12:00 and 14:00-17:00 as the working hours and set a preset dwell time threshold. During these working hours, filter out points where the dwell time of mobile phone users at the same base station and base stations within a radius of 500 meters exceeds the preset threshold. S2.2: Confirm the validity of the filtered results. First, check the location information of the filtered base stations and filter out the base station locations that are not in the actual geographical environment and are in areas where work activities cannot be carried out; Then, abnormal and missing data are filtered out; finally, based on the planning layout of urban functional areas, areas where functional positioning is inconsistent with work behavior are screened, and the validity of the residence point data in the area is further verified to exclude data that does not conform to residential logic; S2.3: Collect the user's signaling data for more than 10 working days. On the premise that the base station data has passed the validity confirmation, if the user's residence time at the same base station representing the work location during the ten working days is greater than the set threshold, the base station is identified as the user's work location.

5. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 1, characterized in that: The process of reading user data in chronological order, determining the status of track points, extracting the trip OD based on the time intervals of consecutive stop points, recording the stop point locations, and then reading the next user data and repeating the above operations includes the following steps: S3.1: Read the current user's mobile phone signaling data in the order of event triggering time; S3.2: Determine the state of the track point. First, set the initial state and determine that each user's initial track point is in the stop state. Then calculate the spatial distance between the track point at the current time t and the previous time t-1. Determine the state of the track point at the current time based on the track point state at the previous time and the comparison result of the current distance with the corresponding distance threshold. The track point status is divided into two states: stay and move. When the previous moment was stay and the distance is greater than the distance threshold, the current moment is determined to be stay and the position is the same as the previous moment; when the distance is greater than the distance threshold, the current moment is determined to be temporarily move; when the previous moment was move and the distance is less than the distance threshold, it is determined that both the previous moment and the current moment are stay and the stay position is different from the previous one; when the distance is greater than the distance threshold, the previous moment is determined to be move and the current moment is temporarily move. S3.3: Propose a trip OD based on the time interval of consecutive stay points. First, calculate the maximum time interval between consecutive static stay points in an area to distinguish different travel behaviors and stay states of users and determine the start and end time and location of the trip. Then, data screening and processing are performed. When the maximum time difference is less than the set time threshold, the stop points in this sequence are judged to be short-term trips and are eliminated. When the maximum time difference is greater than the time threshold, cluster analysis is performed on the consecutive stop points. The start time of the stop point set is determined to be the arrival time of the previous trip, and the end time of the stop point set is determined to be the departure time of the current trip. Finally, the base station with the highest frequency of stop points is used as the starting and ending points of the user's trip. Otherwise, the data group is eliminated and the location information of the stop points is recorded. S3.4: After completing the above operations on the current user's data, read the next user's data and repeat the steps of reading data, determining the status of trajectory points, and extracting travel OD, so as to analyze the data of all target users and generate a trajectory point recognition result table. The trajectory point recognition result table consists of user ID, timestamp, stay field, stay point location and travel OD information, so as to obtain travel OD information of residents in the entire city and specific areas.

6. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 1, characterized in that: The data of fixed equipment with unchanged cleaning location is used to identify the permanent population based on the criterion that the daily stay time exceeds 7 hours and the proportion of days is ≥ 60%, including the following steps: S4.1: Arrange the mobile phone signaling data in chronological order and continuously observe the information of each data point. Set a period of seven working days. If the location of a device remains unchanged during this period, it is preliminarily determined to be fixed device data. If the location of the device remains unchanged for two periods, it is also determined to be fixed device data and the data is cleaned and removed. S4.2: The identification threshold for permanent residents is a daily stay of more than 7 hours per day, with a proportion of days exceeding 60%. This threshold is based on research on residents' daily activity patterns and analysis of actual data, and is used to distinguish permanent residents from other individuals. S4.3: For the cleaned mobile phone signaling data, count the length of stay of each user on different dates and calculate the proportion of days with a daily stay time of more than 7 hours to the total number of days; if this proportion is ≥ 60%, the corresponding user is identified as a permanent resident; in actual operation, use the traversal and judgment algorithm to traverse all user signaling data, complete the screening and marking of qualified users, and realize the identification of permanent residents.

7. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 1, characterized in that: After the permanent population is identified, the non-resident population is divided into transit persons and mobile persons according to the length of stay standard, including the following steps: S5.1: After the permanent resident population has been identified, the remaining population will be identified as non-resident population for the next step of classification; S5.2: The non-resident population is classified based on the length of stay. All non-residents whose stay on a particular date is less than three hours are classified as transit persons, and the rest of the non-resident population are classified as migrants. S5.3: For each individual in the non-resident population, check their stay time information on all appearance dates; if the stay time on all appearance dates is less than 3 hours, classify them as transit personnel; otherwise, classify them as mobile personnel; in actual operation, use traversal and judgment algorithms to traverse and compare the stay time data of all non-resident populations to complete the classification.

8. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 5, characterized in that: Calculating the number of daily trips of a single user based on the state change of the stay field in the trajectory point recognition result table to obtain the total number of daily trips and the average number of daily trips of the user includes the following steps: S6.1: Determine the state of the track point based on the stay field in the track point recognition result table, where stay = 1 indicates stationary and stay = 0 indicates moving; S6.2: For a single user, the process of stay changing from 1 to 0 and then back to 1 represents a completed trip. Therefore, the number of trips per day for a user j can be calculated by counting the number of changes in the stay value. The formula is as follows: ； in, represents the number of trips of user j, n represents the total number of information data of user j, represents the state of the i-th trajectory point of user j, represents the state of the i+1th trajectory point of user j, The absolute value of the difference between the states of two adjacent trajectory points is calculated to determine whether the state of the trajectory point has changed. According to the above formula, the total number of daily trips and the average number of daily trips for all populations can be calculated as follows: ； ； Where m represents the number of valid users.

9. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 5, characterized in that: The method of calculating the travel distance using the determined stop point locations includes first calculating the average travel distance of a single user in a day, and then calculating the total daily, daily average, and trip average travel distances of all users. The method includes the following steps: S7.1: Based on the obtained user stop locations, calculate the distance between the stop points as the straight-line distance of the trip; for user j, use the distinct statement to obtain the unique value set C of the Lon field. Set C contains the coordinates of all the stop points of user j. Set C can be expressed as C={(Lon i , Lat i )|1≤i≤n}; where n is the number of stay points, Lon represents longitude, and Lat represents latitude; S7.2: Based on the coordinates of the stop points obtained above, calculate the travel distance between user j's stop point i and the stop point i+1 by the formula , the formula is as follows: ； in, Indicates the distance, R is the radius of the earth, which is 6371 kilometers, Lon i and Lat i are the longitude and latitude of the i-th stop point, Lon i+1 and Lat i+1 are the longitude and latitude of the i+1th stop point respectively; S7.3: After obtaining the distance between adjacent stop points, calculate the average travel distance of user j in one day according to the formula , the formula is as follows: ； in, represents the average travel distance of user j. Since there are n-1 distances between n stop points, n-1 represents the number of travel segments in a day; S7.4: Add up the travel distances of all users in a day and calculate the total travel distance according to the formula , the formula is as follows: ； in, represents the total daily travel distance, m is the total number of users, represents the sum of m users, represents the sum of the distances between all adjacent stay points of each user j in one day; S7.5: Then divide the total daily travel distance by the total number of users m and calculate the average daily travel distance according to the formula , the formula is as follows: ； in, is the average daily travel distance, represents the total daily travel distance, m is the total number of users; S7.6: Finally, calculate the average travel distance of all users in a day. First, calculate the total number of trips of all users in a day. Let the number of trips for each user j be F j , the total number of trips is , and then calculate the average travel distance of all users in a day through the formula , the formula is as follows: ； in, is the average trip distance, represents the total daily travel distance, The total number of trips for all users in a day. This value reflects the average distance of each trip.

10. The method for analyzing resident travel characteristics based on mobile phone signaling data according to claim 1, characterized in that: The steps of filtering movement point times to form a set, setting time period parameters to filter database records, calculating travel distribution in different time periods, and completing the analysis of residents' travel characteristics in the time dimension include the following steps: S8.1: Distribute travel time into one-hour time periods, count the number of trips occurring within these time periods, and calculate the travel incidence rate for each time period. The travel incidence rates for each time period are linked together in chronological order to form the time distribution of travel. The travel time distribution objectively reflects the temporal distribution of travel demand. S8.2: First, filter the movement point times to form a set. Use the data table as the operation object and Stay=0 as the query condition. Use the select statement to filter out the Timestamp fields of all movement points, which are recorded as set D. Set D is expressed as D={Timestamp i |1≤i≤x}, where x is the number of all moving points; S8.3: Then set two parameters p and q, satisfying (0 < p < q < 24), to define the time interval; the parameters p and q divide the 24 hours of a day into different time periods; S8.4: Based on the obtained set D, the database records are filtered again with Stay = 0 and (p < Timestamp < q) as the condition to obtain all records that occurred in the time period from p to q, which is recorded as set E. Set E is expressed as E = {Timestamp i |1≤i≤z}, where z is the number of elements in set E; S8.5: Finally, calculate the probability H of traveling from time p to time q according to the formula pq , the formula is as follows: ； Among them, 0 < p < q < 24, x is the total number of movement points within a day, and z is the number of movement points generated from moment p to moment q; by calculating the H pq value, the travel occurrence probability at any time period within a day is obtained, and the travel distribution in different time periods is determined accordingly; S8.6: Finally, arrange the travel incidence rates for each time period in chronological order to form the time distribution of travel, thereby completing the analysis of residents' travel characteristics in the time dimension.

Images