The national data portal: data.gov.sg
Launched in 2011 and rebuilt on a new technical stack in 2018, data.gov.sg is Singapore's central repository for government-published datasets. As of April 2026, it hosts over 3,000 datasets from more than 70 public agencies, covering categories from real-time environmental readings to historical HDB resale transaction records.
For smart city applications, the most-accessed datasets on data.gov.sg include: the NEA 2-hour weather forecast (updated every 30 minutes from the network of 65 weather stations); PM2.5 hourly readings by region; the LTA real-time carpark availability feed (covering 1,800+ public carparks); and the OneMap basemap tiles and geocoding API, which provides the authoritative Singapore address reference.
All datasets on data.gov.sg are published under the Singapore Open Data Licence, which permits free use, distribution, and adaptation for both non-commercial and commercial purposes, subject to attribution. The licence is compatible with the Creative Commons Attribution 4.0 licence for practical purposes.
The LTA DataMall
The Land Transport Authority operates a separate developer portal — DataMall — specifically for transport data. Unlike data.gov.sg, DataMall requires free registration and issues API keys, a structure that allows LTA to monitor usage and enforce rate limits on high-frequency real-time feeds.
DataMall's most frequently accessed endpoints, based on LTA's published API call statistics, are:
- Bus arrival times — predicted arrival for all stops on all routes, updated every 30 seconds from onboard GPS units in Singapore's full 5,800-bus fleet.
- Traffic images — live snapshots from 87 expressway and arterial road cameras, updated every 5 minutes.
- Carpark availability — real-time lot counts for HDB and URA carparks, mirroring the data.gov.sg feed at higher update frequency.
- Taxi availability — GPS positions of available taxis, updated every 30 seconds.
- ERP rates — current and scheduled future pricing for all ERP gantries and (under ERP 2.0) zone boundaries.
The Virtual Singapore platform
Virtual Singapore, developed by GovTech in partnership with the French geospatial software firm Dassault Systèmes, is a 3D city model that integrates physical geometry — buildings, terrain, vegetation, underground utilities — with real-time data layers from the sensor networks described above. It is not a public data portal; access is restricted to government agencies and accredited research institutions via a formal application process.
The platform's primary documented uses include: solar irradiance mapping for rooftop photovoltaic feasibility assessment (NParks has used it to identify optimal locations for community solar installations); wind flow modelling in proposed new development sites to assess natural ventilation; and flood inundation simulation for drainage master planning by PUB.
The 3D base model is updated quarterly from aerial LiDAR surveys conducted by the Singapore Land Authority. Real-time sensor data layers are ingested from the central GovTech data repository, with a 15-minute lag from field measurement to availability within Virtual Singapore's analysis environment.
The Moments of Life / LifeSG platform
On the resident-facing end of the data infrastructure, the LifeSG mobile application (formerly Moments of Life) aggregates government service interactions and data into a single interface. From a smart city data perspective, its most relevant feature is the integration of real-time municipal service data: bus and MRT arrival times, park and hawker centre crowd density indicators derived from footfall sensor data, and air quality alerts triggered when PSI readings exceed threshold values.
LifeSG had over 1.5 million registered users as of December 2024 according to the GovTech annual report. The footfall and crowd data visible within the app is derived from anonymised counts from the NEA and NParks sensor networks; individual movement data is not retained.
Data governance and the Public Sector Data Governance Framework
Underlying all the platforms above is a governance layer that determines what data can be shared, at what granularity, and with whom. The Public Sector Data Governance Framework, administered by the Smart Nation and Digital Government Group, defines three tiers of data sensitivity: Restricted (government-internal only), Confidential (agency-specific with inter-agency sharing under agreement), and Open (publishable under the Singapore Open Data Licence).
Environmental sensor readings at city-district aggregation (not individual node level) are classified Open. Traffic camera images are classified Confidential, published on DataMall only as still snapshots with a 5-minute delay. Individual vehicle movement records from the ERP system are classified Restricted and are not accessible to external parties.
API reliability and documented service levels
GovTech publishes a service-level commitment for data.gov.sg APIs: 99.5% monthly uptime, with real-time feeds updated within 10% of their stated refresh interval 95% of the time. DataMall's stated SLA is slightly lower at 99% uptime, reflecting the higher update frequency and greater infrastructure demands of the transport feeds.
A 2023 independent analysis by a team from the Singapore Management University, published in the journal Data & Policy, benchmarked Singapore's open data infrastructure against 18 comparable city-state and small-nation open data programmes. Singapore ranked first on API stability and third on data freshness (median time between sensor reading and API availability), behind Copenhagen and Amsterdam. The primary gap identified was dataset documentation depth — specifically, the absence of machine-readable data dictionaries for 38% of datasets on data.gov.sg at the time of analysis.
What is not available
For researchers working with Singapore's urban data, several gaps are worth noting. The Virtual Singapore 3D model is not publicly accessible. Real-time individual-node sensor readings (as opposed to district aggregates) require a research data access agreement with the relevant agency. CCTV footage from the island's network of approximately 90,000 public cameras is not available in any form through the open data infrastructure. And while ERP transaction volumes are published in aggregate, route-level vehicle movement data remains restricted.
The SNDGO has indicated that a revised data access framework, extending limited access to individual-node sensor data for accredited researchers, is under development, with consultation expected in 2025. Details on the framework's scope and application process had not been published as of this article's last review date.