Method and device for identifying cross-city commuting users based on mobile user signaling data

Abstract

The present invention discloses a method and device for identifying cross-city commuting users based on mobile user signaling data, including: based on the daytime period and nighttime period of each cycle, using the mobile user signaling data of the target area to obtain the daily working users and daily resident users of the cycle; according to the nighttime signaling data of the daily working users of the cycle and the daytime signaling data of the daily resident users of the cycle, respectively obtaining the nighttime signaling disappearance users and daytime signaling disappearance users of the cycle; using the number of times that all mobile users become the nighttime signaling disappearance users or daytime signaling disappearance users within a set time period to obtain the identification result of cross-city commuting users. The present invention mines cross-city commuting users based on original signaling data and uses the Spark computing framework for analysis and processing, which has high reliability and high efficiency and can be used for regional population supervision.

Description

Technical Field

[0001] The present invention belongs to the field of information technology, and particularly relates to a method and device for identifying cross-city commuting users based on mobile user signaling data. Background Art

[0002] With the development of urbanization and transportation means, as well as the radiation and driving effect of big cities on the surrounding economy, the phenomenon that the place of residence and the place of work belong to different regions is becoming more and more obvious. Many people choose to commute across cities, which brings certain difficulties in population management, such as in the fields of epidemic control, public safety management, urban traffic planning, etc.

[0003] Traditional cross-city commuting population supervision can be obtained through checkpoints such as highway toll stations, railway stations, airports, etc., but it faces problems such as scattered and one-sided data, and cannot reflect the cross-city commuting population situation of the whole region. The rapid progress of mobile communication technology provides good technical support for cross-city commuting population supervision. The signaling services of mobile users are frequent, including telephone and SMS services, Internet access services, location update services, etc. High-frequency communication behaviors generate a large amount of spatio-temporal information data, which can well reflect the movement trajectories of mobile users to a certain extent. Statistical data of the Ministry of Industry and Information Technology in October 2020 shows that there are currently 1.6 billion mobile phone users in China, of which 1.29 billion are 4G users and 1.33 billion are mobile Internet users; there are 39.044 million mobile users in Beijing, and 32.122 million 4G users.

[0004] The signaling types in signaling data include tunnel update, service request, handover, location update, tunnel establishment, tunnel deletion, attachment, call originating start, power on, receiving SMS, call originating end, detachment, power off, call terminating start, sending SMS, call terminating end, receiving MMS, etc. Statistical results show that the proportion of power on and power off signaling data of mobile users in Beijing in September 2020 accounted for about 0.25% of the signaling data. Therefore, the mobile phones of mobile users are basically in the power on state, and it is reasonable to judge whether a user is in the city based on signaling data.

[0005] At present, the research on commuting behavior based on mobile signaling data mainly focuses on intra-city commuting, and the population analysis of cross-city commuting has not been carried out yet. In addition, the existing commuting analysis adopts the method of first feature extraction and then clustering analysis, which is not feasible when dealing with the signaling data of millions of people in a city, because generally, the existing computing clusters are difficult to support the massive operations required for clustering analysis. Summary of the Invention

[0006] To fill the gap in the existing cross-city commuting analysis and in combination with the actual situation, the embodiments of the present invention provide a method and device for identifying cross-city commuting users based on mobile user signaling data. By using the data of working users, resident users and mobile user signaling data, the target users meeting the conditions are identified and merged into cross-city commuting users, and the identification results can objectively and accurately reflect the statistical data of cross-city commuting users in the current city.

[0007] To achieve the above object, the present invention adopts the following technical solutions:

[0008] A method for identifying cross-city commuting users based on mobile user signaling data, the steps of which include:

[0009] 1) Based on the daytime period and nighttime period of each cycle, the daytime signaling data and nighttime signaling data of each mobile user in the mobile user signaling data of the target area are respectively extracted to obtain the daily working users of this cycle and the daily resident users of this cycle;

[0010] 2) According to the nighttime signaling data of the daily working users of this cycle and the daytime signaling data of the daily resident users of this cycle, the nighttime signaling disappearing users of this cycle and the daytime signaling disappearing users of this cycle are respectively obtained;

[0011] 3) Using the number of times that all mobile users become the nighttime signaling disappearing users or the daytime signaling disappearing users within the set time period, the identification result of cross-city commuting users is obtained.

[0012] Further, the types of mobile user signaling data include: call detail record data, circuit switched domain data and / or packet switched domain data.

[0013] Further, the information of mobile user signaling data includes: user name, timestamp and date.

[0014] Further, the method for analyzing the information of mobile user signaling data includes: the distributed computing framework Spark.

[0015] Further, the daytime signaling data and nighttime signaling data of each mobile user are extracted through the following steps:

[0016] 1) In the mobile user signaling data, all user names and signaling timestamp fields within this cycle are obtained and de-duplicated;

[0017] 2) For the de-duplicated data, the signaling timestamps are sorted by user name to obtain the signaling timestamp sequence of each mobile user in this cycle;

[0018] 3) Based on different time periods, the daytime signaling data and nighttime signaling data of this cycle are extracted from the signaling timestamp sequence of this cycle.

[0019] Further, the users with disappeared periodic night signaling are obtained through the following steps:

[0020] 1) If the periodic night signaling data of a periodic daily working user is empty, it is recorded as an all - type user with disappeared night signaling;

[0021] 2) Among the periodic daily working users excluding the all - type users with disappeared night signaling, if the time stamp of the first signaling data in the periodic night signaling data of a periodic daily working user is later than M hours after the start time of the night period, it is recorded as a beg - type user with disappeared night signaling;

[0022] 3) Among the periodic daily working users excluding the all - type and beg - type users with disappeared night signaling, if the time stamp of the last signaling data in the periodic night signaling data of a periodic daily working user is earlier than M hours before the end time of the night period, it is recorded as an end - type user with disappeared night signaling;

[0023] 4) Among the periodic daily working users excluding the all - type, beg - type, and end - type users with disappeared night signaling, if the time - stamp difference between any two signaling data in the periodic night signaling data of a periodic daily working user is greater than or equal to M hours, it is recorded as a mid - type user with disappeared night signaling;

[0024] 5) By combining the all - type, beg - type, end - type, and mid - type users with disappeared night signaling, the users with disappeared periodic night signaling are obtained.

[0025] Further, the users with disappeared periodic day signaling are obtained through the following steps:

[0026] 1) If the periodic day signaling data of a periodic daily resident user is empty, it is recorded as an all - type user with disappeared day signaling;

[0027] 2) Among the periodic daily resident users excluding the all - type users with disappeared day signaling, if the time stamp of the first signaling data in the periodic day signaling data of a periodic daily resident user is later than N hours after the start time of the day period, it is recorded as a beg - type user with disappeared day signaling;

[0028] 3) Among the periodic daily resident users excluding the all - type and beg - type users with disappeared day signaling, if the time stamp of the last signaling data in the periodic day signaling data of a periodic daily resident user is earlier than N hours before the end time of the day period, it is recorded as an end - type user with disappeared day signaling;

[0029] 4) Among the daily resident users in this cycle after excluding all types of users with disappeared daytime signaling, beg types of users with disappeared daytime signaling, and end types of users with disappeared daytime signaling, if the time stamp difference between any two signaling data in the daytime signaling data of the daily resident users in this cycle is greater than or equal to N hours, they are recorded as mid types of users with disappeared daytime signaling;

[0030] 5) By synthesizing all types of users with disappeared daytime signaling, beg types of users with disappeared daytime signaling, end types of users with disappeared daytime signaling, and mid types of users with disappeared daytime signaling, the users with disappeared daytime signaling in this cycle are obtained.

[0031] Furthermore, the identification results of cross-city commuting users include: monthly commuting users who work in the city and live outside the city, and monthly commuting users who live in the city and work outside the city.

[0032] Furthermore, the monthly commuting users who work in the city and live outside the city, and the monthly commuting users who live in the city and work outside the city are obtained through the following strategies:

[0033] 1) When the number of times any mobile user is identified as a user with disappeared nighttime signaling in the set time period is greater than the threshold, it is considered as a monthly commuting user who works in the city and lives outside the city;

[0034] 2) When the number of times any mobile user is identified as a user with disappeared daytime signaling in the set time period is greater than the threshold, it is considered as a monthly commuting user who lives in the city and works outside the city;

[0035] A storage medium stores a computer program, wherein the computer program is set to execute any of the above methods when running.

[0036] An electronic device includes a memory and a processor. The memory stores a computer program, and the processor is set to run the computer program to execute any of the above methods.

[0037] Compared with the prior art, the present invention has the following advantages:

[0038] 1. The present invention fills the gap in identifying cross-city commuting users based on mobile user signaling data and can be used for regional population supervision;

[0039] 2. The present invention calculates and identifies users with disappeared nighttime and daytime signaling using the signaling data of the entire city's population, and the results are reliable and have practical application reference value;

[0040] 3. The present invention uses the distributed computing framework Spark to analyze and process massive data, and the calculation process is efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0042] Figure 2 It is a schematic diagram of the algorithm flow for identifying users with disappeared night signals in the present invention.

[0043] Figure 3 It is a schematic diagram of the algorithm flow for identifying users with disappeared day signals in the present invention.

[0044] Figure 4 It is a city-wide statistical chart of users with disappeared night signals in the embodiment.

[0045] Figure 5 It is a distribution chart of the proportion of four types of users with disappeared night signals in the embodiment.

[0046] Figure 6 It is a statistical chart of monthly commuting users who work in the city and live outside the city in the embodiment.

[0047] Figure 7 It is a city-wide statistical chart of users with disappeared day signals in the embodiment.

[0048] Figure 8 It is a distribution chart of the proportion of four types of users with disappeared day signals in the embodiment.

[0049] Figure 9 It is a statistical chart of monthly commuting users who live in the city and work outside the city in the embodiment. Detailed implementation manner

[0050] In order to illustrate the technical solution disclosed in the present invention in detail, further elaboration will be made below in combination with the embodiments and the drawings in the specification.

[0051] The method for identifying cross-city commuting users in the present invention includes the following steps:

[0052] (1) Obtain the signaling data of daily working users and daily living users in the target area;

[0053] (2) Screen out users whose night signals have continuously disappeared for M hours on the same day from the daily working users as users with disappeared night signals, and screen out users whose day signals have continuously disappeared for N hours on the same day from the daily living users as users with disappeared day signals;

[0054] (3) Take users whose number of days with disappeared night signals in the current month exceeds the set threshold as monthly commuting users who work in the city and live outside the city; take users whose number of days with disappeared day signals in the current month exceeds the set threshold as monthly commuting users who live in the city and work outside the city.

[0055] Further, the target area described in step (1) is generally selected as a city to be analyzed. The daily working users and daily living users are generally determined by combining data analysis and survey statistics. The mobile user signaling data includes call detail records, circuit switched domain data, and packet switched domain data. The signaling data with invalid usernames is excluded, the username and signaling timestamp fields are selected and de-duplicated, and sorted in chronological order to form a user's daily signaling timestamp sequence.

[0056] The night period described in step (2) is from 21:00 on the same day to 07:00 on the next day. The signaling generated during this period is called night signaling. Four steps are used to determine the users whose night signaling disappears continuously for M hours:

[0057] (2.1a) If a daily working user has no signaling during the night period, it is recorded as an all-class night signaling disappearance user;

[0058] (2.1b) After excluding all-class users, if the first night signaling of a daily working user appears at least M hours after 21:00 on the same day, it is recorded as a beg-class night signaling disappearance user;

[0059] (2.1c) After excluding all, beg-class users, if the last night signaling of a daily working user appears at least M hours before 07:00 on the next day, it is recorded as an end-class night signaling disappearance user;

[0060] (2.1d) After excluding all, beg, end-class users, if the time difference between the timestamps of two consecutive night signalings of a daily working user is greater than or equal to M hours, it is recorded as a mid-class night signaling disappearance user.

[0061] The above four types of user sets are mutually exclusive. Each type of user is calculated step by step and finally merged into the night signaling disappearance users of the current day.

[0062] The day period described in step (2) is from 07:00 to 19:00 on the same day. The signaling generated during this period is called day signaling. Four steps are used to determine the users whose day signaling disappears continuously for N hours:

[0063] (2.2a) If a daily living user has no signaling during the day period, it is recorded as an all-class day signaling disappearance user;

[0064] (2.2b) After excluding all-class users, if the first day signaling of a daily living user appears at least N hours after 07:00 on the same day, it is recorded as a beg-class day signaling disappearance user;

[0065] (2.2c) After excluding all, beg-class users, if the last day signaling of a daily living user appears at least N hours before 19:00 on the same day, it is recorded as an end-class day signaling disappearance user;

[0066] (2.2d) After excluding users of the all, beg, and end types, if the time difference between the timestamps of the two consecutive signaling messages of a daily resident user during the day is greater than or equal to N hours, then the user is recorded as a mid-type daytime signaling disappearance user.

[0067] The above four types of user sets are mutually exclusive. Calculate each type of user step by step, and finally merge them into the daytime signaling disappearance users of the current day.

[0068] Step (3) Based on the data of nighttime signaling disappearance users and daytime signaling disappearance users for one month, use the aggregation method for screening: users who meet the condition that the number of days of nighttime signaling disappearance in the current month exceeds the set threshold are regarded as monthly commuting users who work in the city and live outside the city; users who meet the condition that the number of days of daytime signaling disappearance in the current month exceeds the set threshold are regarded as monthly commuting users who live in the city and work outside the city.

[0069] The following describes the specific process.

[0070] (1) Obtain the signaling data of daily working users and the signaling data of daily resident users in the target area.

[0071] In this embodiment, the signaling data used is the signaling data of Beijing mobile users that has been encrypted for sensitive information. The user data is the daily working users and daily resident users in Beijing determined through analysis. The amount of data in the original data source is huge and stored in Hive, and the analysis and processing of the data use the distributed computing framework Spark.

[0072] Based on the daily working user data, daily resident user data, and the mobile user signaling data of the current day and the next day in the target area, associate the daily working user data with the signaling data to obtain the daily working user signaling data, and associate the daily resident user data with the signaling data to obtain the daily resident user signaling data.

[0073] The time range of the signaling data and user data is September 2020, and the spatial range is Beijing. The signaling data packet includes call detail record data, circuit switched domain data, and packet switched domain data. The information included is username, timestamp, location, signaling type, related identifier, etc. The signaling database only records the user data that generates signaling services within the scope of Beijing. According to the needs of the present invention, three effective data information of date, encrypted user number, and signaling timestamp are extracted. The signaling timestamps of a certain mobile user in one day are shown in Table 1.

[0074] Table 1 Signaling Timestamps of a Certain Mobile User in One Day

[0075]

[0076]

[0077]

[0078] The information structure of daily working users includes date, administrative region, street, and encrypted user number, as shown in Table 2. The information structure of daily resident users is the same as that of daily working users.

[0079] Table 2 Information Structure of Working Users

[0080] Date Administrative Region Street User Number 20200922 Haidian District Huayuan Road Street 86106XSJCU3D1 20200922 Changping District Cuicun Town 86106ASGDL27G 20200922 Fengtai District Lugouqiao Street 86106XA2E8CYX 20200922 Chaoyang District Wangsi Camp Township 86106D3CIVYVU 20200922 Tongzhou District Yuqiao Street 8610636F34F9F 20200922 Shunyi District Gaoliying Town 86106SDW7SXC

[0081] (2) For daily working users, screen out users whose signaling continuously disappears for M hours during the night. The night period is from 21:00 on the same day to 07:00 the next day. The signaling generated during this period is called night signaling. Determine users whose night signaling continuously disappears for M hours in 4 steps. In the embodiment, M = 6 is taken.

[0082] (2.1a) If a daily working user does not generate signaling during the night period, it is recorded as an all - type night - signaling - disappearing user;

[0083] (2.1b) After excluding all - type users, if the first night signaling of a daily working user appears at least M hours after 21:00 on the same day, it is recorded as a beg - type night - signaling - disappearing user;

[0084] (2.1c) After excluding all - type and beg - type users, if the last night signaling of a daily working user appears at least M hours before 07:00 the next day, it is recorded as an end - type night - signaling - disappearing user;

[0085] (2.1d) After excluding all - type, beg - type, and end - type users, if the time - stamp difference between the two adjacent night signalings of a daily working user is greater than or equal to M hours, it is recorded as a mid - type night - signaling - disappearing user.

[0086] The above four types of user sets are mutually exclusive. Calculate each type of user step by step and finally merge them into the night - signaling - disappearing users of the same day. The result of analysis and processing based on the user data of Beijing in September 2020 is as Figures 4 - 5 shown, where Figure 4 is the statistical chart of night - signaling - disappearing users, Figure 5 is the distribution chart of the proportion of the four types of night - signaling - disappearing users. It can be seen that users without signaling at night and without signaling in the second half of the night account for 70%.

[0087] (3) For daily resident users, screen out users whose signaling continuously disappears for N hours during the day. The day period is from 07:00 to 19:00 on the same day. The signaling generated during this period is called day signaling. Determine users whose day signaling continuously disappears for N hours in 4 steps. In the embodiment, N = 6 is taken.

[0088] (2.2a) If the daily resident users do not generate signaling during the daytime period, they are recorded as all - type daytime signaling disappearance users.

[0089] (2.2b) After excluding all - type users, if the first daytime signaling of the daily resident users appears at least N hours after 07:00 on the same day, they are recorded as beg - type daytime signaling disappearance users.

[0090] (2.2c) After excluding all - type and beg - type users, if the last daytime signaling of the daily resident users appears at least N hours before 19:00 on the same day, they are recorded as end - type daytime signaling disappearance users.

[0091] (2.2d) After excluding all - type, beg - type, and end - type users, if the time - stamp difference between the two adjacent daytime signalings of the daily resident users is greater than or equal to N hours, they are recorded as mid - type daytime signaling disappearance users.

[0092] The above four types of user sets are mutually exclusive. Calculate each type of user step by step, and finally merge them into the daytime signaling disappearance users of the current day. The results of analysis and processing based on the user data of Beijing in September 2020 are as Figures 7 - 8 shown, where Figure 7 is the statistical chart of daytime signaling disappearance users, Figure 8 is the distribution chart of the proportions of the four types of daytime signaling disappearance users. It can be seen that more than 80% of the daytime signaling disappearance users generate signaling during the daytime, and the signaling disappears continuously for more than 6 hours.

[0093] (4) Take the users whose number of days of nighttime signaling disappearance in the current month exceeds the set threshold as monthly commuting users who work in the city and live outside the city, and take the users whose number of days of daytime signaling disappearance in the current month exceeds the set threshold as monthly commuting users who live in the city and work outside the city.

[0094] Take the users in Beijing in September 2020 whose number of days of nighttime signaling disappearance exceeds 10 days as monthly commuting users who work in the city and live outside the city in September, and take the users whose number of days of daytime signaling disappearance exceeds 10 days as monthly commuting users who live in the city and work outside the city in September. Figure 6 is the statistical chart of monthly commuting users who work in the city and live outside the city, which is close to the ranking trend of the population numbers of each administrative region in Beijing. Figure 9 is the statistical chart of monthly commuting users who live in the city and work outside the city, which is also close to the ranking trend of the population numbers of each administrative region in Beijing.

[0095] The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Those skilled in the art can modify the technical solutions of the present invention or make equivalent replacements without departing from the spirit and scope of the present invention. The protection scope of the present invention shall be subject to what is described in the claims.

Claims

1. A method for identifying cross - city commuting users based on mobile user signaling data, the steps of which include: 1) Based on the daytime period and nighttime period of each cycle, extract the daytime signaling data and nighttime signaling data of each mobile user in the mobile user signaling data of the target area for each cycle, so as to obtain the daily working users and daily living users of this cycle; among them, the information of the mobile user signaling data includes: user name, timestamp, and date; 2) According to the nighttime signaling data of the daily working users of this cycle and the daytime signaling data of the daily living users of this cycle, obtain the nighttime signaling disappearance users and daytime signaling disappearance users of this cycle respectively; among them, obtaining the nighttime signaling disappearance users according to the nighttime signaling data of the daily working users of this cycle includes: If the nighttime signaling data of a daily working user of this cycle is empty, it is recorded as an all - type nighttime signaling disappearance user; Among the daily working users excluding all - type nighttime signaling disappearance users, if the timestamp of the first signaling data in the nighttime signaling data of a daily working user of this cycle is later than M hours after the start time of the nighttime period, it is recorded as a beg - type nighttime signaling disappearance user; Among the daily working users excluding all - type nighttime signaling disappearance users and beg - type nighttime signaling disappearance users, if the timestamp of the last signaling data in the nighttime signaling data of a daily working user of this cycle is earlier than M hours before the end time of the nighttime period, it is recorded as an end - type nighttime signaling disappearance user; Among the daily working users excluding all - type nighttime signaling disappearance users, beg - type nighttime signaling disappearance users, and end - type nighttime signaling disappearance users, if the time - stamp difference between any two signaling data in the nighttime signaling data of the daily working user of this cycle is greater than or equal to M hours, it is recorded as a mid - type nighttime signaling disappearance user; Integrate all - type nighttime signaling disappearance users, beg - type nighttime signaling disappearance users, end - type nighttime signaling disappearance users, and mid - type nighttime signaling disappearance users to obtain the nighttime signaling disappearance users of this cycle; 3) Use the number of times that all mobile users become nighttime signaling disappearance users or daytime signaling disappearance users within a set time period to obtain the identification result of cross - city commuting users.

2. The method according to claim 1, characterized in that the types of mobile user signaling data include: call detail data, circuit - switched domain data, and / or packet - switched domain data.

3. The method according to claim 1, characterized in that the method for analyzing mobile user signaling data information includes: the distributed computing framework Spark.

4. The method according to claim 1, characterized in that extract the daytime signaling data and nighttime signaling data of each mobile user through the following steps: 1) In the mobile user signaling data, obtain all user names and signaling timestamp fields within this cycle and remove duplicates; 2) Sort the signaling timestamps by user name for the data after duplicate removal to obtain the signaling timestamp sequence of each mobile user for this cycle; 3) Based on different time periods, extract the daytime signaling data and the nighttime signaling data of this cycle from the cycle signaling timestamp sequence.

5. The method according to claim 1, characterized in that the daytime signaling disappearing users of this cycle are obtained through the following steps: 1) If the daytime signaling data of a daytime resident user of this cycle is empty, it is recorded as an all - type daytime signaling disappearing user; 2) Among the daytime resident users of this cycle excluding the all - type daytime signaling disappearing users, if the timestamp of the first signaling data in the daytime signaling data of a daytime resident user of this cycle is later than N hours after the start time of the daytime period, it is recorded as a beg - type daytime signaling disappearing user; 3) Among the daytime resident users of this cycle excluding the all - type daytime signaling disappearing users and the beg - type daytime signaling disappearing users, if the timestamp of the last signaling data in the daytime signaling data of a daytime resident user of this cycle is earlier than N hours before the end time of the daytime period, it is recorded as an end - type daytime signaling disappearing user; 4) Among the daytime resident users of this cycle excluding the all - type daytime signaling disappearing users, the beg - type daytime signaling disappearing users, and the end - type daytime signaling disappearing users, if the time - stamp difference between any two signaling data in the daytime signaling data of a daytime resident user of this cycle is greater than or equal to N hours, it is recorded as a mid - type daytime signaling disappearing user; 5) Combine the all - type daytime signaling disappearing users, the beg - type daytime signaling disappearing users, the end - type daytime signaling disappearing users, and the mid - type daytime signaling disappearing users to obtain the daytime signaling disappearing users of this cycle.

6. The method according to claim 1, characterized in that The cross - city commuting user identification results include: monthly commuting users who work in the city and live outside the city, and monthly commuting users who live in the city and work outside the city; The monthly commuting users who work in the city and live outside the city, and the monthly commuting users who live in the city and work outside the city are obtained through the following strategies: 1) When the number of times any mobile user is identified as a nighttime signaling disappearing user of this cycle within a set time period is greater than the threshold, it is considered a monthly commuting user who works in the city and lives outside the city; 2) When the number of times any mobile user is identified as a daytime signaling disappearing user of this cycle within a set time period is greater than the threshold, it is considered a monthly commuting user who lives in the city and works outside the city.

7. A storage medium, in which a computer program is stored, wherein the computer program is set to execute the method according to any one of claims 1 - 6 when running.

8. An electronic device, including a memory and a processor, a computer program is stored in the memory, and the processor is set to run the computer program to execute the method according to any one of claims 1 - 6.

Images