Sensing the Chinese Diaspora: How Mobile Apps Can Provide Insights Into Global Migration Flows

Minhui Xue, Xin Yuan, Heather Lee, Keith W. Ross

This paper presents a novel, "passive" methodology for tracking global migration and diaspora patterns by leveraging the location-based services (LBS) of country-centric mobile apps. Using WeChat as a primary case study, the researchers demonstrate how "teleporting" to global cities via GPS hacking can reveal fine-grained data on ethnic communities that traditional census and survey methods often miss.

TL;DR

Researchers used GPS hacking and automated screen scraping on WeChat to "sense" the Chinese diaspora across 32 global cities. By analyzing the "People Nearby" feature, they generated real-time insights into migration density and cultural assimilation that are more granular and current than traditional census data.

What Problem Does This Paper Address?

Studying the "diaspora"—the dispersion of people from their original homeland—is notoriously difficult. Traditional methods like national censuses are infrequent (often every 10 years), while surveys are expensive and suffer from low response rates.

Existing digital methods also have flaws: IP-based geolocation is often too coarse (identifying a country but not a neighborhood), and Twitter-based studies are biased toward specific, often highly-educated demographics. The authors sought a way to collect high-resolution, time-series data on specific ethnic and cultural groups without the high costs or lag time of traditional sociology.

Methods

The authors developed a three-tiered automated pipeline to collect data from "country-centric" apps (apps primarily used by residents of one country, like WeChat for China, KakaoTalk for South Korea, or VK for Russia).

  1. GPS Hacking: Using software to spoof a smartphone’s GPS coordinates, the researchers "placed" virtual phones at the city halls of 32 target cities across five continents.
  2. Task Automation: They used Appium, a mobile automation tool, to script interactions within the WeChat app. The script would open the "People Nearby" feature, scroll through the list of users within a 2,000-meter radius, and take screenshots of profiles and "Moments" (social feeds).
  3. Data Extraction (OCR): They processed over 75,000 screenshots using Tesseract-OCR to extract usernames, distances, and status messages in both Chinese and Latin characters.
  4. Classification: To ensure they were only counting the Chinese diaspora, they used a combination of automated detection (checking for Chinese characters) and manual verification by students to filter out non-ethnic Chinese users.

Datasets

The study produced a unique dataset by taking measurements on four consecutive Saturday afternoons in 2016.

  • Scope: 32 cities including Prato (Italy), Suva (Fiji), San Francisco (USA), and Johannesburg (South Africa).
  • Scale: 6,308 distinct WeChat users were identified within the 2km probe zones, 3,205 of whom were confirmed as ethnic Chinese.
  • Comparison: The researchers compared their findings against Table I, which lists other potential "sensors" like LINE (Japan), Zalo (Vietnam), and JDate (Jewish diaspora).

Key Insights and Ideas

  • Apps as Cultural Beacons: Because WeChat is the "everything app" for Chinese nationals, its usage abroad serves as a high-fidelity proxy for the presence of the diaspora, particularly for the 18–35 age demographic.
  • The "Prato" Outlier: The data identified Prato, Italy, as having the highest concentration of young Chinese diaspora—more than double the concentration of the next highest city. This aligns with journalistic evidence of Prato's massive Chinese-led garment industry, validating the mobile-sensing approach.
  • Language as an Assimilation Metric: By analyzing the language used in status updates, the researchers could estimate integration. For instance, they found high levels of local language usage (assimilation) in cities like Johannesburg and Berlin, but very low levels in Prato and Bucharest, suggesting these communities remain more culturally insular.
  • Consistency: The relative density of users remained remarkably stable across different weeks, suggesting that even a single "snapshot" of LBS data can provide a reliable estimate of a city's diaspora population.

Key Results

  • Density Rankings: After normalizing for city population, the highest concentrations of the Chinese diaspora were found in Prato (Italy), Suva (Fiji), Antwerp (Belgium), and Lisbon (Portugal).
  • Market Penetration: In most studied cities, over 70% of the people using the "People Nearby" feature were ethnic Chinese. However, in some regions (like India and South America), WeChat's aggressive marketing to locals meant that ethnic Chinese users were only a small fraction of the total LBS pool.
  • Validation: The findings corroborated known migration trends, such as the recent surge of Chinese investment and tourism in Fiji.

Limitations and Future Work

The authors acknowledge several constraints:

  • Spatial Limitation: The study only probed a 2,000-meter radius around city halls. This misses "ethnoburbs" or ethnic enclaves located on the outskirts of major cities.
  • Demographic Bias: The "People Nearby" feature is predominantly used by younger people (ages 18–35), meaning the data does not accurately represent older generations of immigrants.
  • Manual Effort: While the data collection was automated, high-accuracy labeling still required manual intervention to distinguish ethnic Chinese users from locals using the app.
  • Future Directions: The authors suggest scaling the system to use multiple "probes" per city to create a heat map of diaspora density and applying the methodology to other country-centric apps like KakaoTalk or VK.

Summary updated

Abstract

Many countries today have "country-centric mobile apps" which are mobile apps that are primarily used by residents of a specific country. Many of these country-centric apps also include a location-based service which takes advantage of the smartphone's API access to the smartphone's current GPS location. In this paper, we investigate how such country-centric apps with location-based services can be employed to study the diaspora associated with ethnic and cultural groups. Our methodology combines GPS hacking, automated task tools for mobile phones, and OCR to generate migration statistics for diaspora. As a case study, we apply our methodology to WeChat, an enormously popular app within China and among ethnic Chinese worldwide. Using WeChat, we collect data about the Chinese diaspora in 32 cities. The combined data provides interesting insights to the modern Chinese diaspora and how it has changed in recent years.
BibTeX
@inproceedings{xue2019sensing,
  author = {Minhui Xue and Xin Yuan and Heather Lee and Keith W. Ross},
  title = {{Sensing the Chinese Diaspora: How Mobile Apps Can Provide Insights Into Global Migration Flows}},
  booktitle = {Workshops @ ICDM},
  year = {2019},
  url = {https://openreview.net/forum?id=TNOzd8ejz8v6},
}
Figure