Transparency

Data methodology

This page explains where the data shown in the app comes from, how it is transformed, how often it is refreshed, and what it does not guarantee.

Last updated: July 28, 2026

Principles Geographic coverage Detailed sources by layer Quality and freshness Freshness schema by source Known limits Official references (map layers)Versioning

Principles

Repères produces no data of its own: the app aggregates datasets published by public bodies, transit agencies, and technical services, normalizes formats, exposes map layers, and computes itineraries. Three rules guide every integration: credit the source on the relevant record, show the real age of the data rather than an approximate “live” label, and visibly flag a source that is down instead of showing a silent blank. A value that is wrong upstream is displayed as-is, with its publication date: we never correct data on a producer's behalf.

Geographic coverage

The app covers 57 Canadian cities grouped into 53 zones, across Quebec, Ontario, Alberta, British Columbia, Manitoba, and Nova Scotia. Coverage is not uniform: each layer exists where an open source is published for that territory. A “transit-first” city (Sherbrooke, Rimouski, most British Columbia systems) may expose only its transit network, while Montreal and Toronto have dozens of municipal layers. The zone selector only shows layers actually available around you.

Detailed sources by layer

Public transit (65 networks)

Greater Montreal: STM (stops, lines, service status, bus and metro realtime) through STM open data, attribution required; exo (trains and buses), RTL, STL, and REM through their open GTFS feeds. Rest of Quebec: RTC (Quebec City), STS (Sherbrooke), Citébus (Rimouski), Transport Saint-Jean, Sorel-Tracy, and STQ ferries. Ontario: TTC, GO Transit and UP Express, OC Transpo, MiWay, Brampton Transit, York Region Transit, Durham Region Transit, Burlington Transit, London, Hamilton, Waterloo, and Barrie. West and Atlantic: TransLink (Vancouver), 35 BC Transit systems, Edmonton Transit Service, Calgary Transit, Winnipeg Transit, Halifax Transit, and Fort McMurray. Scheduled times come from static GTFS feeds, mirrored on our backend to reduce load on upstream sources; realtime relies on GTFS-RT or proprietary feeds where the agency publishes them. Not every network broadcasts realtime: in that case scheduled times are shown and labelled as such.

Shared mobility

Bike share: BIXI (Montreal), Bike Share Toronto, and other station-based systems, all through their official GBFS feed, for realtime bike and dock availability. Shared e-scooters: operator GBFS feeds, in cities where the service is deployed. Car share: Communauto stations and Flex vehicles, used with source attribution and in line with the Communauto licence. EV charging stations: municipal datasets and OpenStreetMap depending on the city.

Municipal open data portals

Five portals feed the neighbourhood layers. Données Montréal: parking, taxis, EV chargers, REV, cycle tracks, bike racks, murals, public art, book boxes, public pianos, ecocentres, libraries, cultural venues, fountains, parks, food establishments, events, sports and leisure, pools, recreational facilities, ice rinks, winter sites, ski conditions, waste collection, and snow clearing. Données Québec: Laval, Longueuil, Repentigny, and Gatineau (parks, sports facilities, wetlands, waste collection). Toronto Open Data: parks and recreation facilities, ice rinks, drinking fountains, public WiFi, public art, off-leash areas, bicycle parking, EV chargers, libraries, and DineSafe inspections. Ottawa Open Data: parks, greenspace, and recreation facilities. OpenStreetMap fills the gaps where no municipal dataset exists. Licences are mostly CC-BY 4.0; producer attribution is preserved on every record.

Traffic, public safety, and environment

Roadwork and closures, traffic incidents, speed cameras, and traffic cameras: Données Montréal, Données Québec, and the 511 road network depending on the layer. Traffic flow: TomTom, under a daily request quota. Wildfires: SOPFEU for Quebec and the Ministry of Natural Resources for Ontario. Marine traffic: AIS positions. Flight tracking, International Space Station passes, and aurora forecasts: specialised public services, for information only.

Weather and air quality

Weather, hourly forecasts, and nowcast: Open-Meteo. Air quality: Montreal's RSQA through Données Québec (hourly index and stations), falling back to Open-Meteo when unavailable. Air quality health index and water levels: Environment and Climate Change Canada. Official weather alerts: ECCC.

Power outages

Outage areas (polygons) and incidents from the official Hydro-Québec Open Data / Info-pannes source, with a transparent fallback when it is unavailable. Coverage matches the territory served by Hydro-Québec.

Basemap and imagery

The map uses MapLibre GL JS. The vector basemap is built on Protomaps and OpenStreetMap, served as self-hosted PMTiles and regenerated automatically. Satellite and Hybrid modes use Esri World Imagery; Hybrid overlays Protomaps labels.

Itineraries, search, and geocoding

Walking, cycling, and driving: self-hosted Valhalla engine, whose graph is built from OpenStreetMap extracts clipped around every covered city (Quebec, Ontario, Alberta, British Columbia, Manitoba). Public transit: OpenTripPlanner, whose graph currently contains only the STM GTFS feed and the Montreal OpenStreetMap extract — transit itineraries are therefore available in Montreal only, while the other modes work across every covered city. Address search uses Geoapify when configured, then falls back to Nominatim and OpenStreetMap; reverse geocoding goes through the backend to reduce load on upstream sources.

Quality and freshness

Every layer exposes a status (current, stale, unavailable) and a last-updated time. If a source goes down, the last valid data may stay on screen alongside its status, to avoid an artificial drop to zero that would read as a genuine absence.

Freshness schema by source

Indicative target cadences per source family. They may vary with upstream availability and the network involved.

Source	Layers	Target cadence	Fallback
STM (GTFS, GTFS-RT, service status)	Lines, stops, next departures, service status, vehicle positions	20s to 90s depending on endpoint	Last valid data kept on screen, with a visible “stale” or “unavailable” status.
Regional networks (GTFS and GTFS-RT)	Stops, lines, and schedules for the 64 non-STM networks; realtime where the agency publishes it	30s for realtime, daily for static	Last valid data kept on screen, with a visible “stale” or “unavailable” status.
Bike share and e-scooters (GBFS)	Stations, available bikes and docks, free-floating vehicles	30s to 60s	Last valid data kept on screen, with a visible “stale” or “unavailable” status.
Données Montréal	Parking, taxis, chargers, REV, cycle tracks, bike racks, murals, book boxes, cultural venues, fountains, parks, restaurants, events, sports and leisure, pools, recreational facilities, waste collection, snow clearing	Minutes to daily depending on the dataset	Last valid data kept on screen, with a visible “stale” or “unavailable” status.
Municipal portals outside Montreal	Neighbourhood layers for Toronto, Ottawa, Laval, Longueuil, Repentigny, and Gatineau; OpenStreetMap as a complement	Daily to weekly depending on the portal	Last valid data kept on screen, with a visible “stale” or “unavailable” status.
Traffic and public safety	Roadwork and closures, incidents, traffic flow, cameras, speed cameras, wildfires	2 min to 15 min depending on the layer	Last valid data kept on screen, with a visible “stale” or “unavailable” status.
Open-Meteo, RSQA, and ECCC	Weather, air quality, official alerts, water levels	≈ 1h	Last valid data kept on screen, with a visible “stale” or “unavailable” status.
Hydro-Québec Open Data / Info-pannes	Outage areas and incidents	≈ 30s to 5 min depending on the source	Last valid data kept on screen, with a visible “stale” or “unavailable” status.

STM (GTFS, GTFS-RT, service status)

Layers: Lines, stops, next departures, service status, vehicle positions
Target cadence: 20s to 90s depending on endpoint
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Regional networks (GTFS and GTFS-RT)

Layers: Stops, lines, and schedules for the 64 non-STM networks; realtime where the agency publishes it
Target cadence: 30s for realtime, daily for static
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Bike share and e-scooters (GBFS)

Layers: Stations, available bikes and docks, free-floating vehicles
Target cadence: 30s to 60s
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Données Montréal

Layers: Parking, taxis, chargers, REV, cycle tracks, bike racks, murals, book boxes, cultural venues, fountains, parks, restaurants, events, sports and leisure, pools, recreational facilities, waste collection, snow clearing
Target cadence: Minutes to daily depending on the dataset
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Municipal portals outside Montreal

Layers: Neighbourhood layers for Toronto, Ottawa, Laval, Longueuil, Repentigny, and Gatineau; OpenStreetMap as a complement
Target cadence: Daily to weekly depending on the portal
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Traffic and public safety

Layers: Roadwork and closures, incidents, traffic flow, cameras, speed cameras, wildfires
Target cadence: 2 min to 15 min depending on the layer
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Open-Meteo, RSQA, and ECCC

Layers: Weather, air quality, official alerts, water levels
Target cadence: ≈ 1h
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Hydro-Québec Open Data / Info-pannes

Layers: Outage areas and incidents
Target cadence: ≈ 30s to 5 min depending on the source
Fallback: Last valid data kept on screen, with a visible “stale” or “unavailable” status.

Known limits

Transit itineraries are computed in Montreal only; elsewhere, walking, cycling, and driving are the available modes. Not every network publishes realtime data: on those, the times shown are scheduled. Some agencies only publish the upcoming season in their GTFS calendar, which can leave a stop without schedules for several weeks. Vehicle positions depend on GPS coverage and may be several minutes old in tunnels. Seasonal layers (ice rinks, winter sites, ski conditions) remain visible out of season with a matching label. Finally, publication delays, geographic rounding, gaps between scheduled and realtime data, and upstream outages can affect accuracy: any critical use must be validated through the official channels of the organisation concerned.

Official references (map layers)

For every integration, the official source is shown on the detail record when available, and on the service monitoring page. Main portals and services:

Versioning

Methodology schema version: 2026.07. This page is reviewed on every significant coverage extension; user-visible changes are published on the updates page.

Public updates: Public updates