Exploring the Lives of Domestic Helpers in Hong Kong Through Geospatial Analysis
Final Project
IA 310 Mapping World Cities
Sidath Chandrasena
1.0 Introduction
In many parts of the world, migrant communities are subdivided into groups, including expatriate workers and migrant laborers. Due to the rise in population without economic growth, imported labor has become a product of globalization across the world, ranging from skilled jobs to housework. This is prevalent particularly in places such as countries in the Middle East as well as Singapore, and Hong Kong. One particularly interesting area is the distribution of live-in domestic helpers, often women, who reside in their employers’ homes, which prevents the formation of migrant enclaves by residence. Many of these workers often face labor exploitation and abuse at the hands of their employers. Through geospatial analysis, I aim to examine the distribution of live-in domestic helpers in Hong Kong, explore where they find community, and their access to migrant support centers.
2.0 Study Area
Hong Kong is a Special Administrative Region (SAR) located in China (Government of Hong Kong, 2024). With a population of 7.54 million people as of 2023 , it is the fourth most densely populated region in the world (Ritchie & Mathieu, 2019). The evolution of the region to a global financial center is attributed to the history of British colonialism, when Hong Kong was ceded to the United Kingdom in 1842 under the Treaty of Nanking following the defeat of the Qing Dynasty in the Opium War (Halliday, 1974). The city remained largely under British control, apart from temporary occupation by Imperial Japan during World War II, before the handover to the Chinese government in 1997 (Johnson, 1986).
In terms of its geography, several aspects have influenced its evolution and identity as a city. In the late 1830s, the island was described by the British as “little more than a barren rock, speckled with a few tiny fishing villages” (Carroll, 1997). The vested interest of the British in Hong Kong’s natural harbor led to heavy investment and the economic growth in the region to transform the city from a fishing village into a global trading hub (Kestell & Meinheit, 1997). During the Japanese occupation from 1941 to 1945, the city was changed into a shipping hub to aid the military effort in the Pacific region (Man & Lun, 2014), trade and commerce was halted, and the conditions became unbearable for citizens resulting in depopulation of the territory, as citizens finding the conditions unbearable fled to Mainland China and Macau (Swee-Hock & Kin, 1975). The Japanese occupation and Chinese Civil War resulted in altering of the landscape by removing surrounding forest for fuel production, a period which “was a disaster for Hong Kong forestry” (Corlett, 1999, p. 96). Mass migration from Mainland China during major historical events such as the Chinese Civil War and the foundation of the People’s Republic of China in 1949 also altered the city’s landscape (Madokoro, 2012). Developed land makes up less than a quarter of total land area in Hong Kong and forested land makes up around half of the total land area, and subsequently the rapid population growth led to a housing crisis and land reclamation projects, with around 6% of the city made of reclaimed land, 27% of the population residing on this land, and 70% of the city’s business activities taking place on this land (Ng, 2020). The lack of public housing has resulted in the poorest residents of the city residing in cage homes, rooftop slums, and led to the formation of enclaves such as the notorious Kowloon Walled City (Yu, 2020). The rapid rise in population necessitated an efficient transportation system, resulting in the development of the Mass Transit Railway (MTR), which has become one of the most heavily utilized public transport networks in the world (Lo et al., 2008). Another important aspect that has affected the landscape of Hong Kong has been the changing identity of the city under the “One Country, Two Systems” constitutional agreement between the UK and China in 1997, guaranteeing autonomy for the city for at least 50 years (So, 2011). Since 1997, there have been several instances of urban planning and landscape changes to foster integration between Hong Kong and the mainland, such as the investment in high-speed rail infrastructure from Hong Kong to Guangzhou and Shenzhen, the Hong Kong-Zhuhai-Macau bridge, and developments within the border of Hong Kong and Shenzhen (Yuen & Cheng, 2020).
Several observers have considered Hong Kong to be a prominent city in the global world-city network due to its present-day role as a leading financial hub (Chu, 2008). The importance of this term to the city is seen through its own self-branding efforts from 1997 as “Asia’s World City” (Shen, 2010). The city is currently in a state of major change due to the changing political landscape and income inequality between Chinese and non-Chinese migrants. Although a study by Lee et al. (2007) found that unlike other major cities, minority groups are not clustered around a single area, Chinese migrants were found to reside in largely low-income areas while non-Chinese migrants reside primarily in affluent areas. Upon further analysis, it was found that non-Chinese migrants compose of skilled expatriates from high-income countries such as Japan and the United States and live-in domestic helpers, who comprise of primarily women from Southeast Asia, namely the Philippines, Indonesia, and Thailand (Lee et al., 2007). Female domestic helpers (FDHs) provided a solution in the 1980s to Hong Kong women, who struggled between fulfilling their Confucian beliefs of managing household duties and filial piety and joining the workforce with increasing opportunities due to the city’s economic boom (Lee et al., 2007).
Given that FDHs make up around five percent of Hong Kong’s total population, I am curious to understand more about their geographical distribution in the city. Although they may not be residentially represented due to living in their employers’ home, I would like to use population distribution to identify which districts have a higher concentration of Filipinos and Indonesians. After identifying these districts, I would like to identify physical spaces where FDHs are able to find community for social, religious or cultural purposes, and their proximity to support centers given the rampant abuse and exploitation that takes place.
3.0 Methods and Data
For analysis on migrant groups in Hong Kong, I consulted the Esri China Hong Kong website, which contained geospatial datasets on each province in Hong Kong. These datasets were originally obtained from the Hong Kong government’s public and open data portal, which contains a lot of extensive datasets on various sectors such as climate and weather, education, and city management at utilities.
The most relevant datasets were the Population Distribution by Ethnicity, which was available from 2004 until 2024. This contains data on the population of a variety of ethnic groups, including Filipino, Indonesians and Thais. However, this data is from the population census, which excludes female domestic helpers (FDHs). There is data in a CSV format for FDHs, but it is much less detailed, with simply the year and total population. I decided to use the population distribution data to gain insight into where these migrant communities are situated (e.g. restaurant owners, etc) and then use that information to learn more about where FDHs may be inclined to visit on their rest days (Sundays).
First, install the relevant packages and set the working directory:
require(tidyverse)
Loading required package: tidyverse
── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr 1.1.4 ✔ readr 2.1.5
✔ forcats 1.0.0 ✔ stringr 1.5.1
✔ ggplot2 3.5.1 ✔ tibble 3.2.1
✔ lubridate 1.9.3 ✔ tidyr 1.3.1
✔ purrr 1.0.2
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag() masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
require(sf)
Loading required package: sf
Linking to GEOS 3.8.0, GDAL 3.0.4, PROJ 6.3.1; sf_use_s2() is TRUE
Warning in CPL_read_ogr(dsn, layer, query, as.character(options), quiet, : GDAL
Message 1: This version of GeoPackage user_version=0x000028A0 (10400, v1.4.0)
on '/cloud/project/Population_distribution_by_Ethnicity_in_2001.gpkg' may only
be partially supported
Reading layer `Ethnicity' from data source
`/cloud/project/Population_distribution_by_Ethnicity_in_2001.gpkg'
using driver `GPKG'
Simple feature collection with 18 features and 8 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 12672060 ymin: 2529946 xmax: 12739620 ymax: 2579129
Projected CRS: WGS 84 / Pseudo-Mercator
Warning in CPL_read_ogr(dsn, layer, query, as.character(options), quiet, : GDAL
Message 1: This version of GeoPackage user_version=0x000028A0 (10400, v1.4.0)
on '/cloud/project/Population_distribution_by_Ethnicity_in_2006.gpkg' may only
be partially supported
Reading layer `Ethnicity_2006' from data source
`/cloud/project/Population_distribution_by_Ethnicity_in_2006.gpkg'
using driver `GPKG'
Simple feature collection with 18 features and 8 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 12672060 ymin: 2529946 xmax: 12739620 ymax: 2579129
Projected CRS: WGS 84 / Pseudo-Mercator
Warning in CPL_read_ogr(dsn, layer, query, as.character(options), quiet, : GDAL
Message 1: This version of GeoPackage user_version=0x000028A0 (10400, v1.4.0)
on '/cloud/project/Population_distribution_by_Ethnicity_in_2011.gpkg' may only
be partially supported
Reading layer `Ethnicity_2011' from data source
`/cloud/project/Population_distribution_by_Ethnicity_in_2011.gpkg'
using driver `GPKG'
Simple feature collection with 18 features and 24 fields
Geometry type: MULTIPOLYGON
Dimension: XY
Bounding box: xmin: 12672060 ymin: 2529946 xmax: 12739620 ymax: 2579129
Projected CRS: WGS 84 / Pseudo-Mercator
From 2011 onwards, there is the inclusion of much more specific categories including Chinese, Indonesian, Filipino, White, Indian, Pakistani, Nepalese, Japanese, Thai, Other Asian, and Others. Given that I am interested in studying FDHs in particular, I will be looking at the Filipino, Indonesian, and Thai columns, as those three groups make up the majority of this category. Therefore, I will be using the 2011, 2016, and 2021 data for this project.
To visualize and understand where the different districts are located, here is a plot of the districts with labels for each district:
ggplot(data = pop_2021) +geom_sf(aes(fill=District)) +guides(fill=FALSE)+geom_label_repel(data = pop_2021, aes(label=District, geometry = geometry), stat='sf_coordinates', size=2)+labs(title="Map of Hong Kong by District")
Warning: The `<scale>` argument of `guides()` cannot be `FALSE`. Use "none" instead as
of ggplot2 3.3.4.
Warning in st_point_on_surface.sfc(sf::st_zm(x)): st_point_on_surface may not
give correct results for longitude/latitude data
I observed that the 2021 columns had slightly different column names, with the English and Chinese names merged, so I renamed the columns to make it consistent:
To calculate proportions for different ethnic groups, I needed a Total column. The 2021 dataset had one, but the others did not, so I made new columns for 2011 and 2016.
pop_2011 <- pop_2011 %>%mutate(Total = Chinese + Indonesian + Filipino + White + Indian + Pakistani + Nepalese + Japanese + Thai + Other_Asian + Others)pop_2016 <- pop_2016 %>%mutate(Total = Chinese + Indonesian + Filipino + White + Indian + Pakistani + Nepalese + Japanese + Thai + Other_Asian + Others)pop_2021 <- pop_2021 %>%rename("Total"="Total_population___總人口")
3.1.1 Filipinos in Hong Kong
First, I will be changing the pop_2011 and pop_2016 datasets to calculate the proportion of Filipinos.
Comparing the percentage of Filipinos in Hong Kong from 2011 to 2021:
ggplot(data = pop_2011, aes(fill = prop_filipino)) +geom_sf()+scale_fill_continuous(limits=c(0, 10))+labs(title="Filipinos in Hong Kong - 2021")
ggplot(data = pop_2016, aes(fill = prop_filipino)) +geom_sf()+scale_fill_continuous(limits=c(0, 10))+labs(title="Filipinos in Hong Kong - 2016")
ggplot(data = pop_2021, aes(fill = prop_filipino)) +geom_sf()+scale_fill_continuous(limits=c(0, 10)) +labs(title="Filipinos in Hong Kong - 2021")
From the above figure, it appears that the overall proportion of Filipinos as part of the total population has increased over time, with significant amounts in Yau Tsim Mong and Wan Chai.
To see the spread within the Filipino population in Hong Kong rather than Total Population, I will also calculate the location quotient and map the results.
The location quotient reveals that there is a reasonable amount of variation in where the Filipino population resides. A higher distribution of the Filipino population resides in Central and Western, Wan Chai, Islands, Southern, Kowloon as well as Eastern, Sai Kung, and Yau Tsim Mong to a lesser extent.
3.1.2 Indonesians in Hong Kong
Indonesians make up a smaller portion of the total population, but I still repeated the same steps to see if anything insightful would be shown.
ggplot(data = pop_2011, aes(fill = prop_indo)) +geom_sf()+scale_fill_continuous(limits=c(0, 3.5))+labs(title="Indonesians in Hong Kong - 2011")
ggplot(data = pop_2016, aes(fill = prop_indo)) +geom_sf()+scale_fill_continuous(limits=c(0, 3.5))+labs(title="Indonesians in Hong Kong - 2016")
ggplot(data = pop_2021, aes(fill = prop_indo)) +geom_sf()+scale_fill_continuous(limits=c(0, 3.5)) +labs(title="Indonesians in Hong Kong - 2021")
Note that this time, the bounds are between 0 to 3 percent compared to 0 to 10 percent for Filipinos. Compared to the map for Filipinos, it is a lot less varied, ranging from 1-3% across areas. Like the Filipinos, it appears that Wan Chai is where the Indonesian population makes up the greatest percentage out of the total population.
I then calculated the location quotient to see the distribution within the Indonesian population in Hong Kong:
The location quotient yielded similar results to the previous plot, with most of the Indonesian population being located in Wan Chai, Central and Western, Eastern, Southern, Yau Tsim Mong, Kowloon City, Tsuen Wan, and Tai Po compared to other areas that had a location quotient less than 1.
3.1.3 Thais in Hong Kong
The 2021 data puts Thais in the “Others” column, so only the 2011 and 2016 data was used.
ggplot(data = pop_2011, aes(fill = prop_thai)) +geom_sf()+scale_fill_continuous(limits=c(0, 1))+labs(title="Thais in Hong Kong - 2011")
ggplot(data = pop_2016, aes(fill = prop_thai)) +geom_sf()+scale_fill_continuous(limits=c(0, 1))+labs(title="Thais in Hong Kong - 2016")
The smaller size of the Thai community means that calculating proportions out of the total population provides quite limited insight, so the location quotient will be calculated again to get more helpful information about where they are distributed:
This result is quite different from the Filipino and Indonesian location quotient results. The most popular districts in 2011 were Wan Chai, followed by Central & Western, Southern, Islands, and Kowloon City. However, by 2016, Islands became the most popular district, as well as Central & Western, Yau Tsim Mong, and Kowloon with Wan Chai going down to fifth place.
3.2 Identifying FDH Congregation Sites
To add to the information I gained from looking at my data, I researched online and found information from a mixture of formal and informal sources about places where Filipinos, Indonesians, and Thais congregate in Hong Kong to see if they matched my findings and contextualized them. I was able to identify at least 5 points and created KML files for Filipino and Thai congregation points by creating and exporting layers for each on Google MyMaps.
Figure 1
FDHs gathering in Victoria Park, Hong Kong (Guatri, 2012)
3.2.1 Filipino FDH Congregation Sites
I found that the World-Wide House arcade in Central is frequented by FDHs, with many shops run by Filipinos (Chen, 2021). This reminded me of Lucky Plaza in Singapore, where domestic helpers to go to transfer money, get Filipino food, and ship boxes of food and clothes to their loved ones back home. I also learned about the mass gathering of Filipino women on Sundays in Central for their one day off a week, where they use phone booths, eat Filipino food, and consume products from Filipino specialty shops, a ritual that has taken place since the early 1980s (Moss, 2017). Due to the area being a “dead public space” during the weekends without office employees, FDHs congregate there with plastic mats and tents (Chen, 2021). Five sites I identified included:
World-Wide House
Status Square
St Joseph’s Church
Former Hong Kong Court of Final Appeal Building
Bank of America Tower
3.2.2 Indonesian FDH Congregation Sites
I learned that Indonesian FDHs gather primarily in Victoria Park in Wan Chai (Guatri, 2022). To save money, they sit outside on cardboard and tarpaulins and share food with one another. I also learned that Causeway Bay located on the border of Eastern and Wan Chai is termed “Little Indonesia” (Verebes, 2024). Five sites I identified included:
Victoria Park
Fuk Wah Street Rest Garden
Tamar Park
Kowloon Mosque And Islamic Centre
Yuen Long Park
3.2.3 Thai FDH Congregation Sites
Kowloon City is described online as Hong Kong’s “Little Thailand”, with dozens of Thai restaurants and grocery stores, which is also where domestic workers visit during Sundays. The community was affected by the facelift to the area, causing rent prices to increase, motivating Thais to move to places such as Sham Shui Po and Wong Tai Sin (Lin, 2022). A minority visits Thai temples or church, however the low religiosity is reflective of the lack of accessibility on public transport and the fact that monks visit Kowloon City on Sundays (Hewison, 2004). I was unable to find any information on locations where Thai domestic workers met, which was expected as they represent a much smaller number of FDHs.
3.3 Proximity to Support Centers
Given that transportation was mentioned as a challenge for domestic helpers, who try to save as much money as possible, I wanted to see how accessible it may be for domestic helpers to reach a migrant support center on Sundays when they traverse these areas.
Reading layer `SSCEC' from data source `/cloud/project/HKG_support_centers.gpkg' using driver `GPKG'
Simple feature collection with 8 features and 24 fields
Geometry type: MULTIPOINT
Dimension: XY
Bounding box: xmin: 811643 ymin: 815314 xmax: 841067 ymax: 833787.6
Projected CRS: Hong Kong 1980 Grid System
ggplot(data = pop_2016) +geom_sf(data = pop_2016) +guides(fill=FALSE)+geom_sf(data = support_centers, color="blue")+geom_sf(data = filipino_locations, color="red")+labs(title="Filipino Congregation Sites & Migrant Support Centers in Hong Kong")
As seen above, the Filipino congregation locations I found were mostly concentrated between Yau Tsim Mong and Wan Chai.
Let’s see if the support centers are within walking distance. Let’s say around 1 kilometer a as as buffer.
filipino_location_buffer <-st_buffer(filipino_locations, dist=1500)#filipino_location_buffer_dis <- filipino_location_buffer %>%#summarise()ggplot(data = pop_2016) +geom_sf(data = pop_2016) +guides(fill=FALSE)+geom_sf(data = support_centers, color="blue")+geom_sf(data = filipino_location_buffer, color="red")+labs(title="Filipino Congregation Sites & Migrant Support Centers in Hong Kong")
Visually, it appears that the dissolved buffer falls just short of one of the support centers. In the case that I had a bigger collection of Filipino congregation locations, I would perform a spatial join here.
The result confirms that the Wan Chai support center is just over 1km of distance away, and the Wan Chai support center came up only once I increased the buffer to 1.5 km.
3.3.1 Indonesian FDH Congregation Sites Proximity to Support Centers
For Indonesians, the data points I found were much more scattered across the city:
The result reveals that the support centers are within walking distance of two congregation spots, which is an insightful result.
4.0 Results and Discussion
The results from this analysis are quite insightful into looking at the lives of domestic helpers in Hong Kong. Wan Chai appears to be a place where both Filipinos and Indonesians reside, and while researching about congregation spots, I noticed that the place is known for its bars and entertainment scene, which may attract international crowds and cause it to be more cosmopolitan. When looking at changes in demographics of Filipinos, Indonesians, and Thais, there does appear to have been some changes in the distribution of each group, which may have been due to the Hong Kong protests and changing property prices which may have forced communities to move to places with cheaper rent prices. While it is difficult to determine where exactly live-in domestic helpers reside as the population distribution data excludes these helpers, the analysis helps to see where their respective cultural communities reside, which would attract these helpers on Sundays when they congregate to avoid the isolation they face at their work.
Although Hong Kong is known for its efficient and well-utilized transportation systems by locals, this is a different story for domestic helpers, who try to save as much money as they can to send back home. This explains their choice to camp together in public places, such as parks and train stations. This is also why the results of the support center analysis are important, as many of these workers face abuse and exploitation at their jobs, and without easy access to transport and time aside from Sundays, it may be important that these support centers are available within walking distance.
This analysis has several limitations due to the scope of the project and several improvements can be made. If data is collected in the future on the population distribution of domestic workers, this may help to identify how convenient it is for these workers to get to the outlined congregation points. Additionally, although I only identified five congregation points based on literature and news articles, more could be added with more extensive research to see whether more spots lie closer to support centers. There could also be more research done to identify the congregation spots and experiences of smaller live-in domestic helper communities from Nepal and Sri Lanka. Furthermore, there are several other NGOs and organizations centered around helping foreign domestic workers which could also be incorporated into the analysis.
5.0 Conclusion
This project helps provide insight into the different migrant communities that go largely unnoticed in Hong Kong. The first section of the report revealed some of the changes that took place within the Filipino, Indonesian, and Thai migrant communities over time as various factors such as the Hong Kong protests and a housing crisis took place, causing the distribution of these communities across the city to change. These findings helped identify areas within the city that had higher proportions of these communities, which motivated domestic helpers from these countries to congregate with their fellow domestic helpers in these areas. By identifying a number of these congregation locations, the accessibility to these support centers was measured. This analysis could be useful for the establishment of future support centers and directing other resources at supporting these communities.
6.0 Reference List
Carroll, J. M. (1997). Colonialism and collaboration: Chinese subjects and the making of British Hong Kong. China Information, 12(1-2), 12-33.
Chu, Y. W. (2008). Deconstructing the global city: Unravelling the linkages that underlie Hong Kong’s world city status. Urban Studies, 45(8), 1625-1646.
Corlett, R. T. (1999). Environmental forestry in Hong Kong: 1871–1997. Forest ecology and management, 116(1-3), 93-105.
Halliday, J. (1974). Hong Kong: Britain’s Chinese colony. New Left Review, 87(88), 91-112.
Hewison, K. (2004). Thai migrant workers in Hong Kong. Journal of Contemporary Asia, 34(3), 318-335.
Johnson, G. E. (1986). 1997 and after: Will Hong Kong survive? A personal view. Pacific Affairs, 59(2), 237-254.
Kestell, J & Meinheit, H. (1997, August). Hong Kong: From Fishing Village to Financial Center”. Library of Congress Information Bulletin. Library of Congress. https://www.loc.gov/loc/lcib/9708/hongkong.html
Lo, H. K., Tang, S., & Wang, D. Z. (2008). Managing the accessibility on mass public transit: The case of Hong Kong. Journal of Transport and Land Use, 1(2), 23-49.
Madokoro, L. (2012). Borders transformed: Sovereign concerns, population movements and the making of territorial frontiers in Hong Kong, 1949–1967. Journal of Refugee Studies, 25(3), 407-427.
Shen, S. (2010). Re-branding without re-developing: constraints of Hong Kong’s ‘Asia’s World City’ brand (1997–2007). The Pacific Review, 23(2), 203-224.
So, A. Y. (2011). “One country, two systems” and Hong Kong-China national integration: A crisis-transformation perspective. Journal of Contemporary Asia, 41(1), 99-116.
Swee-Hock, S., & Kin, C. W. (1975). Population growth and redistribution in Hong Kong, 1841-1975. Southeast Asian journal of social science, 4(1), 123-131.
Yuen, S., & Cheng, E. W. (2020). Between high autonomy and sovereign control in a subnational island jurisdiction: The paradox of Hong Kong under ‘One Country, Two Systems’. Island Studies Journal, 15(1), 131-150.
Yu, J. (2020). Living on the Skyline: Rooftop Housing in Seoul, Hong Kong, and Taipei (Doctoral dissertation).
