CausewayTraffic.sg Logo
CausewayTraffic.sg

Data Methodology & Sources

How CausewayTraffic.sg calculates traffic estimates.

Updated 2026-09-09

1. Public Camera Feeds

Our "Live Traffic Grid" aggregates publicly available CCTV snapshots from the Land Transport Authority (LTA) and OneMotoring. These images are refreshed every 1–5 minutes depending on source availability. We do not own or operate these cameras.

2. Estimation Logic

Our traffic estimates and route recommendations are derived from a hybrid model:

  • Visual Density: Analysis of vehicle queue lengths visible in camera frames (e.g., if queue extends past a specific lamp post).
  • Live Route Estimates: Current route-stage estimates from our Firestore status feed, with separate SG→MY and MY→SG reads for Woodlands and Tuas where available.
  • Historical Trending: Comparison against average travel times for the specific day of week and hour (e.g., Saturday 10 AM historical average).
  • Third-Party Validation: Cross-referenced with Google Maps API travel time data for verification.

If an upstream camera trend feed is delayed, we avoid treating that camera layer as freshly confirmed and fall back to the route estimate, official context, and conservative wording.

3. Status Labels and What They Mean

Every crossing we describe is banded from the minutes we show you, and from nothing else — never from a different figure, and never from the “busier than usual” ratio. The boundaries are the ones our data pipeline uses throughout:

  • Smooth: 20 minutes or less.
  • Moderate: more than 20 and up to 45 minutes.
  • Heavy: more than 45 and up to 75 minutes.
  • Jammed: more than 75 minutes.
  • Updating: nobody measured it. This is not a fifth level of traffic and it is never a substitute for Smooth.

The “busier than usual” ratio compares an hour against its own average rather than telling you how long you will sit there, so we report it beside the estimate as context and never use it as the estimate.

An absent measurement is rendered as an absence. We do not print zero minutes, and we do not fall back to a reassuring word, when a figure is missing or stale: a crossing nobody has measured is not a quiet one.

Our SmartCommute app can still show a different word for the same lane in the same minute. It renders a status label written alongside the estimate rather than derived from it, and the two are computed from different versions of the same measurement. We are aligning them; until that is done, treat the minutes as the figure to plan on, on either surface.

4. Editorial Standards

Our written guides are authored by experienced analysts. We prioritize firsthand experience and verify cross-border regulatory updates against official announcements from JPJ (Malaysia), LTA (Singapore), ICA, MOT, MOM, and MOE before publishing when those sources are relevant.