Population Distributions of Age Groups and Their Influencing Factors Based on Mobile Phone Location Data: A Case Study of Beijing, China

Wenlai Wang, Tao Pei, Jie Chen, Ci Song, Xi Wang, Hua Shu, Ting Ma, Yunyan DuSustainability 2019 · DOI

TL;DR

This paper presents a novel framework for generating fine-grained, hourly population distributions for different age groups in urban areas using mobile phone signaling data. By analyzing these distributions alongside Points of Interest (POI) data, the study reveals how various factors like education, business, and accommodation influence where different age demographics congregate, offering valuable insights for urban planning.

What Problem Does This Paper Address?

Urban planning and resource allocation often require a detailed understanding of where different population segments are located at various times of the day. Traditional census data, while valuable, is often static, infrequent, and lacks the granular spatial and temporal resolution needed to capture the dynamic nature of urban populations. This is particularly true when considering different age groups, each with unique needs and activity patterns. Previous studies have estimated overall urban population distributions using big data, but a fine-grained analysis of how specific age demographics distribute themselves and what factors drive these distributions has been a significant gap. This paper aims to bridge this gap by leveraging mobile phone data to map age-specific population distributions and identify the key influencing factors.

Methods

The core of this research lies in a multi-step framework:

  1. Spatiotemporal Location Reconstruction: Mobile phone signaling data, which records base station location, user ID, timestamp, and age, is processed to reconstruct individual trajectories. This involves addressing "station drift" (unrealistic rapid movements) and merging records to create representative hourly locations for each user.
  2. Gridded Population Computation: The study area (Beijing) is divided into 500x500m grid cells. Population density is then calculated for each grid cell by aggregating mobile phone user locations, using Thiessen polygons to distribute base station coverage and accounting for non-residential areas like water bodies.
  3. Validation: The generated mobile phone-based population densities (specifically for nighttime) are validated against census data at the subdistrict level. This step ensures the reliability of the mobile phone data for population distribution estimation.
  4. Influence Factor Analysis: The Geographic Detector method is employed to analyze the influence of various POI categories (accommodation, business, education, entertainment, restaurant, hospital) on the population distributions of different age groups. This method quantifies the explanatory power of each factor.

Datasets

  • Mobile Phone Signaling Data: Obtained from a major communication operator in Beijing, China. This dataset includes user ID, base station longitude/latitude, recording time, and user age. It covers approximately 16 million users (about 62% of the market) for two specific days in December 2015 (a weekday and a weekend).
  • Points of Interest (POI) Data: Collected from a digital mapping company in Beijing. This dataset contains 16 POI types, with six categories extracted for this study: accommodation, business, education, restaurant, entertainment, and hospital.
  • Population Census Data: From the Sixth Population Census of China (November 2010), at the street scale (315 township-level administrative units). This data was used for validation.

Key Insights and Ideas

  • Age Matters: Different age groups exhibit distinct spatial and temporal distribution patterns, driven by different factors. This highlights the need for age-specific urban planning.
  • Time of Day is Crucial: The influence of POI factors shifts significantly between day and night. Accommodation is dominant at night, while business and restaurant factors become more influential during the day.
  • Education for the Young, Accommodation for the Elderly: Education facilities strongly attract early adulthood individuals (18-24), while accommodation is a primary driver for the elderly (60+).
  • Young and Middle-Aged Seek Services: Business, restaurant, and accommodation are key factors for young (25-39) and middle-aged (40-59) populations, suggesting a strong link between these demographics and commercial/service areas.
  • Hospitals Matter for Seniors: The hospital factor has a notable impact on the distribution of the elderly, underscoring their specific healthcare needs.
  • Entertainment is a Minor Player: The entertainment factor shows very little explanatory power across all age groups, suggesting it's not a primary driver of daily population distribution.
  • Validation is Key: The strong correlation between mobile phone user density and census data (except for the youngest age group) validates the use of mobile data for this type of analysis.

Key Results

  • The study successfully generated fine-grained population distributions for four age groups (18-24, 25-39, 40-59, 60+) and the overall population.
  • Validation against census data showed high correlation (r values between 0.9645 and 0.9787, R² values between 0.9302 and 0.9579) for most age groups, except for adolescents (1-14) due to lower mobile phone ownership.
  • Education is the most influential factor for early adulthood (18-24).
  • Business, restaurant, and accommodation are the primary drivers for young (25-39) and middle-aged (40-59) people.
  • Accommodation is the dominant factor at night, with its influence decreasing during the day, while business and restaurant factors increase in importance.
  • Accommodation is the most powerful factor for the elderly (60+), with hospitals also showing a significant effect.
  • In a case study evaluating coffee shop locations, people aged 25-39 showed the strongest positive correlation with coffee shop presence, while those aged 40-59 showed a negative correlation.

Limitations and Future Work

  • Data Granularity: The mobile phone signaling data used is from a single operator, which may not fully represent the entire urban population. Only two days of data were analyzed.
  • POI Factor Weighting: The study assumes equal weighting for all POIs within a category when using the Geographic Detector method. Future work could explore differential weighting.
  • Future Directions: The authors suggest incorporating other data sources, such as land use types, to further refine the analysis of influencing factors and improve the prediction of reliable fine-grained population distributions. They also aim to improve the representation of the entire population by potentially combining data from multiple operators.

Summary updated

Abstract

The fine-grained population distributions of different age groups are crucial for urban planning applications. With the development of information and communication technology (ICT), detailed population data retrieved from various big data sources, especially on a fine scale, have been extensively used for urban planning. However, studies estimating the detailed population distributions of different age groups are still lacking. This study constructs a framework to generate fine-grained population data for different age groups and explores the influence of various factors on the distributions of different age groups. The population is divided into the following four age groups: (1) early adulthood people: 18 ≤ age ≤ 24, (2) young people: 25 ≤ age ≤ 39, (3) middle-aged people: 40 ≤ age ≤ 59, and (4) elderly people: 60 ≤ age. The results indicate that education and accommodation factors have a major influence on the distributions of early adulthood and elderly people, respectively. Business, restaurant, and accommodation factors are the main factors influencing the population distributions of young and middle-aged people. The accommodation factor plays a major controlling role at night, and its explanatory power gradually decreases during the day, while the explanatory powers of the business and restaurant factors increase and become leading factors during the day. Specifically, the hospital factor has a greater effect on the distribution of elderly people. The entertainment factor has very little explanatory power for the population distributions of the different age groups.
BibTeX
@inproceedings{wang2019population,
  author = {Wenlai Wang and Tao Pei and Jie Chen and Ci Song and Xi Wang and Hua Shu and Ting Ma and Yunyan Du},
  title = {{Population Distributions of Age Groups and Their Influencing Factors Based on Mobile Phone Location Data: A Case Study of Beijing, China}},
  booktitle = {Sustainability},
  year = {2019},
  doi = {10.3390/su11247033},
}
Figure