SINGAPORE - Which issues are top of mind for voters in a given constituency going into the next general election? The answer likely depends on that constituency’s demographics.
With the release of the Electoral Boundaries Review Committee report on March 11, significant changes have been made to the battle lines. There have also been population shifts and the emergence of new housing estates since the last general election in 2020, even within constituencies whose boundaries are largely unchanged.
The Straits Times has conducted an analysis of publicly available resident and housing data to calculate an estimate of each constituency’s demographic profile and to better inform our reporting.
ST understands that major political parties here consider similar data when making decisions such as which areas to contest and where to field certain candidates.
We plan to make use of this data in our in-depth constituency reports, starting with one on Punggol GRC, as well as our constituency hub pages, which will be launched closer to the election campaign period. Various follow-up stories on particular constituencies and key election themes may also cite this data.
Here’s how we did it.
Matching official statistics to new electoral boundaries
We first looked at the latest available data organised by geographical distribution from the Singapore Department of Statistics, namely the “Residential Dwellings by Planning Area/Subzone and Type of Dwelling” and “Singapore Residents by Planning Area/Subzone, Single Year of Age and Sex” data sets from June 2024, as well as the “Resident Population by Ethnic Group” data set from the Singapore Census of Population 2020.
These data sets are sorted by URA Master Plan 2019 planning zones and subzones. However, the subzones do not always neatly correspond to electoral boundaries. In some cases, a subzone may be split across multiple electoral divisions.
To map the data to electoral divisions, we used an open-source software called QGIS to overlay the electoral divisions onto the subzones. We then applied a function to slice the subzones along the electoral boundaries and calculate what portions of the subzone are in a given constituency.
Red lines indicate electoral boundaries while coloured shaded areas indicate URA Master Plan 2019 subzones. PHOTO: Screenshot from QGIS software.
Red lines indicate electoral boundaries while coloured shaded areas indicate URA Master Plan 2019 subzones.PHOTO: SCREENSHOT FROM QGIS SOFTWARE
We found that over 80 per cent of the 332 subzones sit entirely, or almost entirely, within a single group representation constituency or single-member constituency, giving us confidence that all of the residents and homes within those subzones are located neatly within the boundaries of those constituencies. For the remaining subzones, we had to make certain assumptions.
In some cases, we assume that the residential population and homes are distributed evenly across the geographical area of the subzone.
Take, for instance, the Clementi North subzone, which consists almost entirely of residential estates and has close to 31,000 residents. The map data shows that about 40 per cent of the land area of Clementi North is in Holland-Bukit Timah GRC while the remaining 60 per cent is in Jurong East-Bukit Batok GRC.
We therefore assigned about 12,400 of the subzone’s residents to Holland-Bukit Timah GRC and about 18,600 to Jurong East-Bukit Batok GRC while keeping the age distribution consistent across both groups. We did the same with the number of homes and the racial mix.
In other cases, we made judgment calls based on observations of the actual geography of the area.
For example, the land area of the Compassvale subzone in Sengkang is split about evenly between Ang Mo Kio GRC and Sengkang GRC. However, the portion in Ang Mo Kio GRC consists mainly of train depots and empty fields. There are some migrant worker dormitories in the area, but these are excluded from the Department of Statistics data.
We therefore ignored the land split and assigned 100 per cent of the homes and residents in the subzone to Sengkang GRC.
Our approach will naturally be more accurate for constituencies where all or most of their subzones are completely within their boundaries, such as Punggol GRC, and less accurate for constituencies that are less clear-cut. Nevertheless, we are confident that the data gives a reasonably accurate indication of the demographic profile of each constituency.
Join ST's WhatsApp Channel and get the latest news and must-reads.
Thanks for sharing!