A method for extracting people living in two cities who commute between the two places on weekends using mobile phone signaling data

Abstract

A method for extracting people living in two cities who commute between two places on weekends using mobile phone signaling data comprises: obtaining a user's daily residence data set, including four fields: user number, date, residence, and city; identifying a residence on weekdays, merging records with the same user number, daily residence, and weekday city, and adding a frequency field; if the frequency divided by the number of weekdays is greater than or equal to 60%, then the record is the weekday residence; identifying a weekend residence, extracting data from the weekend's daily residence data according to the user number and weekend city in the weekend alternative residence; merging records with the same user number and residence, and adding a frequency field; the record with the largest frequency is the user's weekend residence; identifying people living in two cities, connecting the weekday residence table and the weekend residence table through the user number, extracting records with the same user number in both tables, and these users are people living in two cities.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a method for extracting people living in two cities who commute between two places on weekends by using mobile phone signaling data. Background Art

[0002] Dual-city life refers to living and working in one city on weekdays and living and relaxing in another city on weekends, and this is a regular behavior. With the interconnection of high-speed rail and highways between cities, the cost of cross-city travel has decreased, and regional integration and urbanization are becoming increasingly common. More and more people often return to their hometowns nearby to visit relatives and friends on weekends; or work hard in big cities, settle down in nearby small cities, and travel back and forth between the two places on weekends. Identifying this group of people is of great significance to the co-construction and sharing of regional facilities and the allocation of public services.

[0003] As mobile phone signaling data gradually becomes a mature data product, mobile phone signaling data records the activity trajectory of mobile phone users. By mining the rules of long-term spatiotemporal trajectory, it is possible to identify the user's residence, workplace, recreational area, travel destination and other information. There are public algorithms for cross-city commuting, visiting relatives and returning home, and inter-city travel. The difference between dual-city life and cross-city commuting is that you only travel back and forth between the two places on weekends and have two places of residence, rather than going back and forth every day, and the residence and workplace are in two cities. The difference between dual-city life and visiting relatives and returning home is that the frequency of returning home is higher, basically going back and forth every weekend, not just during short holidays and the Spring Festival. The difference between dual-city life and inter-city travel is that the travel destination is the residence of another city, which is fixed, and only travels back and forth on weekends, rather than having multiple destination cities and irregular travel times. There is no public algorithm for dual-city life yet.

[0004] The study of dual-city life is applied to the research and analysis of regional fields, including metropolitan areas, urbanization, regional integration, etc. The scale and proportion of people living in dual cities who travel back and forth between two places on weekends are counted, and the distribution of residences is visualized. This provides decision support for the formulation of regional development policies. However, there is currently no method to extract people living in dual cities using mobile phone signaling. Summary of the invention

[0005] The present invention aims to overcome the above-mentioned disadvantages of the prior art and provides a method for extracting people living in two cities who commute between the two places on weekends by using mobile phone signaling data.

[0006] In order to identify people living in two cities who commute between the two places on weekends, the present invention needs to solve the following problems:

[0007] (1) Using the intermediate results (residence on the same day) of the processed mobile phone signaling data, according to the change pattern of the residence on the same day on weekdays and weekends, and the relationship between the weekday city and the weekend city, the residence on weekdays and weekends is identified.

[0008] (2) Filter out cities where people regularly live on weekends from their daily residences on weekends, and identify their residences in those cities, excluding daily residences in irregular destinations such as business trips and tourism.

[0009] (3) Identify people living in two cities based on the cities where they live on weekdays and weekends, and construct a dataset of their weekday residence and weekend residence.

[0010] A method for extracting people living in two cities who commute between two places on weekends using mobile phone signaling data comprises the following steps:

[0011] Step 1: Get the user's daily residence data set. The daily residence refers to the location where the user spends the longest time from 0:00 to 5:00 on the same day. Each mobile phone user has a unique residence information every day. The user's daily residence data set includes four fields: user number, date, residence, and city.

[0012] Step 11: Extract the residence data set for working days. Working days are legal working days. Extract the residence data set for working days from the user's residence data set based on the date, including four fields: user number, date, residence, and working day city.

[0013] Step 12: Extract the residence data set for the weekend. The weekend is the ordinary weekend except for the May Day, National Day, and Spring Festival holidays. Considering that some people living in two cities may return to their permanent city early on weekdays and live in another city on the last day of the weekend, the data will show that they also live in this city on the second day (the first working day after the weekend). Therefore, the first working day after the weekend is also counted as the weekend. The residence data set for the weekend includes four fields: user number, date, residence, and weekend city.

[0014] Step 2: Identify weekday residence.

[0015] Merge records with the same user ID, residence on the day, and city on weekdays, and add a frequency field (calculate the number of times the same record appears). If the frequency divided by the number of working days is greater than or equal to 60%, then the record is the residence on weekdays. Includes user ID, residence on weekdays, and city on weekdays. Each user's residence on weekdays is unique.

[0016] Step 3: Identify weekend residences.

[0017] Step 31: Filter the residence on the weekend day. Extract the user ID and weekday city from the weekday residence data, connect the user ID and the weekend day residence data set, and connect the weekday city information to the weekend day residence data set. Delete the records where the weekend city and weekday city are the same.

[0018] Step 32: Weekend coding. Code the weekend number for each user, and the same weekend number is the same.

[0019] Step 33: Identify alternative weekend residences. First merge records with the same user ID, weekend city, and weekend number, then merge records with the same user ID and weekend city, and add a frequency field (calculate the number of times the same record appears). If the frequency divided by the number of weekends is greater than or equal to 50%, keep the record, otherwise delete the record.

[0020] Step 34: Identify weekend residence. According to the user ID and weekend city in the weekend alternative residence, extract data from the residence data of the weekend day, including user ID, date, residence, and weekend city. Merge records with the same user ID and residence, and add a frequency field (calculate the number of times the same record appears). The record with the largest frequency is the weekend residence of the user. Get the weekend residence of each user.

[0021] Step 4: Identify people living in two cities. The weekday residence table and the weekend residence table are connected by user ID, and records with the same user ID in both tables are extracted. These users are people living in two cities, and the retained records are the weekday residence-weekend residence dataset of people living in two cities.

[0022] The advantages of the present invention are:

[0023] (1) Identify people who regularly live in other cities on weekends and their places of residence. This can make up for the shortcomings of existing cross-city commuting algorithms, which can only identify people who travel between two cities every day, can only identify people who travel between two cities once a year for visiting relatives, and the lack of regularity in the identification of people for inter-city travel. This can provide data support for the study of the regularity of residents' mobility between cities, such as metropolitan areas, urbanization, and regional integration.

[0024] (2) Using existing mobile phone signaling data products, there is no need to start from the original data. It can be achieved through simple algorithms such as data classification, group counting, table connection, etc., to deeply explore the value of existing data. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0026] The technical solution of the present invention is further described below in conjunction with the accompanying drawings. Taking the residence of a user on that day in October 2021 as an example, the technical solution of the present invention is further introduced.

[0027] A method for extracting people living in two cities who commute between two places on weekends using mobile phone signaling data comprises the following steps:

[0028] (1) Obtain the user's daily residence data set.

[0029] The user's daily residence data set includes four fields: user number, date, residence, and city.

[0030] (1) Extract the residence data set on weekdays.

[0031] Working days are statutory working days. Based on the date, the residence data set of working days is extracted from the residence data set of the user on that day, including four fields: user number, date, residence, and working day city. The sample data processing results are shown in the following table:

[0032] User Number date Place of residence Weekday City 1 20211008 A X 1 20211009 A X 1 20211011 A X 1 …… …… ……

[0033] (2) Extract the residential location dataset for the weekend day.

[0034] Weekends are ordinary weekends except for the May Day, National Day, and Spring Festival holidays. Considering that some people living in two cities may return to their permanent city early on weekdays and live in another city on the last day of the weekend, the data shows that they also live in this city on the second day (the first working day after the weekend). Therefore, the first working day after the weekend is also counted as the weekend. The residence data set on the weekend day includes four fields: user number, date, residence, and weekend city. The sample data processing results are shown in the following table:

[0035] User Number date Place of residence Weekend City 1 20211010 A X 1 20211011 P Y 1 20211016 P Y 1 20211017 P Y 1 20211018 A X 1 20211023 R Z 1 20211024 B X 1 20211025 B X 1 20211030 P Y 1 20211031 Q Y 1 20211101 Q Y

[0036] (2) Identify the place of residence on weekdays.

[0037] Merge records with the same user ID, current residence, and weekday city, and add a frequency field (to count the number of times the same record appears). The sample data processing results are shown in the table below:

[0038] User Number Place of residence Weekday City Frequency 1 A X 12 1 B X 3 1 P Y 1 1 W Z 1

[0039] If the frequency divided by the number of working days is greater than or equal to 60%, then the record is the workday residence. If there is no record that meets the requirements, the calculation stops. User No. 1 has 70.6% of his workday residence on that day A, A is the workday residence of the user, and the city is X. The workday residence data includes user number, workday residence, and workday city. Each user's workday residence is unique.

[0040] (iii) Identify weekend residence.

[0041] (1) Filter the place of residence on the weekend.

[0042] Extract the user number and weekday city from the weekday residence data, connect the user number and the weekend day residence data set, and connect the weekday city information to the weekend day residence data set. If the weekday residence data and the weekend day residence data set do not have the same user number, the calculation stops. As long as there is a user number that is the same, the calculation continues. The sample data processing results are shown in the table below:

[0043]

[0044]

[0045] Delete the records where the weekend city is the same as the weekday city. If all weekend cities and weekday cities are the same, the calculation stops. As long as there is one record left, the calculation continues. The sample data processing results are shown in the table below:

[0046] User Number date Place of residence Weekend City 1 20211011 P Y 1 20211016 P Y 1 20211017 P Y 1 20211023 R Z 1 20211030 P Y 1 20211031 Q Y 1 20211101 Q Y

[0047] (2) Encode the weekend number for each user, and the same weekend number is the same. The sample data processing results are shown in the following table:

[0048] User Number date Place of residence Weekend City Weekend Number 1 20211011 P Y 1 1 20211016 P Y 2 1 20211017 P Y 2 1 20211023 R Z 3 1 20211030 P Y 4 1 20211031 Q Y 4 1 20211101 Q Y 4

[0049] (3) Identify alternative weekend residences.

[0050] First merge the records with the same user ID, weekend city, and weekend number. See the table below for sample data processing:

[0051] User Number Weekend City Weekend Number 1 Y 1 1 Y 2 1 Z 3 1 Y 4

[0052] Then merge the records with the same user ID and weekend city, and add a frequency field (to calculate the number of times the same record appears). The sample data processing results are shown in the table below:

[0053] User Number Weekend City Frequency 1 Y 3 1 Z 1

[0054] If the frequency divided by the number of weekends is greater than or equal to 50%, the record is retained, otherwise the record is deleted. If no frequency is greater than or equal to 50%, the calculation stops. As long as there is a record with a frequency greater than or equal to 50%, the calculation continues. There are 4 weekends in October 2021. The frequency of user No. 1 living in city Y is 3, the frequency of living in city Z is 1, and the frequency of living in city Y is greater than 50%, so this record is retained.

[0055] (4) Identify weekend residence. According to the user ID and weekend city in the weekend alternative residence, extract data from the residence data of the weekend day, including user ID, date, residence, and weekend city. Merge records with the same user ID and residence, and add a frequency field (calculate the number of times the same record appears). The sample data processing results are shown in the table below:

[0056] User Number Place of residence Frequency 1 P 4 1 Q 2

[0057] The record with the highest frequency is the weekend residence of the user. Get the weekend residence of each user. In the above table, the residence P has the highest frequency and is identified as the residence of user No. 1.

[0058] (IV) Identify people living in dual cities.

[0059] The weekday residence table and the weekend residence table are connected by user ID, and the records with the same user ID in both tables are extracted. These users are the people living in two cities, and the retained records are the weekday residence-weekend residence dataset of the people living in two cities. The sample data processing results are shown in the table below:

[0060] User Number Weekday residence Weekend residence 1 A P

[0061] The key technical point of the present invention is to assign the same number to the same weekend (step 32), merge the same user number, weekend city, weekend number, merge the same user number and weekend city, and calculate the frequency of returning to surrounding cities on weekends by merging twice, exclude irregular weekend residences such as business trips and tourism from the residence on the weekend, and identify the city where regular weekend residence is made (step 33). The dual-city living population is judged based on whether the city where regular weekday residence is made and the city where regular weekend residence is made are the same, and a data set of their weekday residence and weekend residence is constructed (step 4).

[0062] The contents described in the implementation of this specification are merely an enumeration of the implementation forms of the inventive concept. The protection scope of the present invention should not be deemed to be limited to the specific forms described. The protection scope of the present invention also includes equivalent technical means that can be conceived by those skilled in the art based on the inventive concept.

Claims

1. A method for extracting people living in two cities who commute between two places on weekends using mobile phone signaling data, comprising the following steps: Step 1: Get the user's daily residence data set; the daily residence refers to the location where the user spends the longest time from 0:00 to 5:00 on the same day. Each mobile phone user has a unique residence information every day. The user's daily residence data set includes four fields: user number, date, residence, and city; Step 11: Extract the residence data set of working days; working days are legal working days; extract the residence data set of working days from the user's residence data set of the day according to the date, including four fields: user number, date, residence, and working day city; Step 12: Extract the residence data set for the weekend day; the weekend is the ordinary weekend except for the May Day, National Day and Spring Festival holidays, and the first working day after the weekend is also counted as the weekend; the residence data set for the weekend day includes four fields: user number, date, residence, and weekend city; Step 2: Identify the place of residence on weekdays; Merge records with the same user number, residence on the day, and city on weekdays, and add a frequency field, that is, calculate the number of times the same record appears; if the frequency divided by the number of working days is greater than or equal to 60%, then the record is the data of residence on weekdays; the said data of residence on weekdays includes user number, residence on weekdays, and city on weekdays; the residence on weekdays of each user is unique; Step 3: Identify weekend residences; Step 31: Filter the residence on the weekend day; extract the user number and weekday city from the weekday residence data, connect the user number and the weekend day residence data set, and connect the weekday city information to the weekend day residence data set; delete the records where the weekend city and the weekday city are the same; Step 32: Weekend coding: coding the weekend number for each user, and the same weekend number is the same; Step 33: Identify alternative weekend residences; first merge records with the same user ID, weekend city, and weekend number, then merge records with the same user ID and weekend city, and add a frequency field, that is, calculate the number of times the same record appears; if the frequency divided by the number of weekends is greater than or equal to 50%, retain the record, otherwise delete the record; Step 34: Identify the weekend residence; extract data from the residence data of the weekend day according to the user ID and weekend city in the weekend alternative residence, including the user ID, date, residence, and weekend city; merge records with the same user ID and residence, and add a frequency field, that is, calculate the number of times the same record appears; the record with the largest frequency is the weekend residence of the user; obtain the weekend residence of each user; Step 4: Identify people living in two cities; connect the weekday residence table and the weekend residence table through the user number, and extract the records with the same user number in both tables. These users are the people living in two cities, and the retained records are the weekday residence-weekend residence dataset of the people living in two cities.

Images