| name | bilingual-handling |
| description | Patterns for navigating French and English content in Montréal's open data.
Field name mappings, search strategies, and output formatting for bilingual use.
/ Stratégies pour naviguer le contenu bilingue dans les données ouvertes de Montréal.
Correspondances de noms de champs, stratégies de recherche et formatage bilingue.
|
| triggers | ["bilingual, French, English, translate, field names, language","bilingue, français, anglais, traduire, noms de champs, langue"] |
Bilingual Data Handling / Gestion des données bilingues
The Reality / La réalité
Montréal's open data is predominantly French. This is not a bug — it reflects the legal and cultural context of Québec. To work with this data effectively, an AI agent must:
- Search in French — dataset titles, descriptions, tags, and most field names are French
- Recognize bilingual fields — some datasets have paired FR/EN columns
- Present results in the user's language — regardless of the source language
Les données ouvertes de Montréal sont principalement en français. Pour travailler efficacement, un agent IA doit chercher en français, reconnaître les champs bilingues et présenter les résultats dans la langue de l'utilisateur.
Pattern 1: Search Term Translation / Traduction des termes de recherche
When a user asks in English, translate to French before searching:
Municipal Infrastructure
| English | French (search term) |
|---|
| tree(s) | arbre(s) |
| road, street | rue, route, chaussée |
| bridge | pont |
| water main | aqueduc, conduite d'eau |
| building | bâtiment |
| sidewalk | trottoir |
| bike path | piste cyclable, réseau cyclable |
| park | parc |
| fire station | caserne (de pompiers) |
| library | bibliothèque |
| pool, swimming | piscine |
| arena, rink | aréna, patinoire |
Government & Services
| English | French (search term) |
|---|
| budget | budget |
| contract | contrat |
| permit | permis |
| election | élection |
| voting | vote, scrutin |
| borough | arrondissement |
| council | conseil |
| bylaw, regulation | règlement |
| complaint, request | plainte, requête |
| snow removal | déneigement |
| garbage, waste | déchet, collecte, ordure |
Environment
| English | French (search term) |
|---|
| air quality | qualité de l'air |
| water quality | qualité de l'eau |
| green space | espace vert |
| canopy | canopée |
| garden | jardin |
| contaminated soil | terrain contaminé |
| heat island | îlot de chaleur |
Transportation
| English | French (search term) |
|---|
| bus | autobus, bus |
| metro, subway | métro |
| transit | transport (en commun) |
| bike sharing | vélopartage, vélo-partage |
| parking | stationnement |
| road closure | entrave, fermeture |
| construction | travaux |
| traffic | circulation |
| accident, collision | accident, collision |
Pattern 2: Bilingual Field Recognition / Reconnaissance des champs bilingues
Some datasets include explicit bilingual field pairs. The common patterns:
Suffix pattern: _fr / _ang or _en
Essence_fr → "Érable argenté"
Essence_ang → "Silver Maple"
Prefix pattern: NOM_FR / NOM_AN
NOM_FR → "Parc La Fontaine"
NOM_AN → "La Fontaine Park"
No bilingual pair (most common)
Most fields are French-only. The field name itself is in French:
ARROND_NOM → "Ville-Marie" (borough name, FR only)
Emplacement → "Banquette gazonnée" (location type, FR only)
Rue → "Rue Sainte-Catherine" (street name, bilingual by nature)
How to handle French-only fields
-
Proper nouns (street names, park names, borough names) — use as-is in both languages. "Rue Sainte-Catherine" stays "Rue Sainte-Catherine" in English output.
-
Category values — translate when presenting to English users:
INV_TYPE: "H" = "Hors rue" (English: "Off-street")
INV_TYPE: "R" = "Rue" (English: "Street")
-
Data dictionary terms — translate column descriptions:
DHP = "Diamètre à Hauteur de Poitrine"
→ "Diameter at Breast Height" (DBH, standard forestry term)
Pattern 3: Borough Names / Noms d'arrondissements
Borough names are proper nouns — they are the same in both languages but must be matched exactly (including accents and special characters):
| Code | Full Name |
|---|
| AHU | Ahuntsic-Cartierville |
| ANJ | Anjou |
| CDN | Côte-des-Neiges–Notre-Dame-de-Grâce |
| LAC | Lachine |
| LAS | LaSalle |
| PLA | Le Plateau-Mont-Royal |
| LSO | Le Sud-Ouest |
| IBI | L'Île-Bizard–Sainte-Geneviève |
| MHM | Mercier–Hochelaga-Maisonneuve |
| MTN | Montréal-Nord |
| OUT | Outremont |
| PRF | Pierrefonds-Roxboro |
| RDP | Rivière-des-Prairies–Pointe-aux-Trembles |
| RPP | Rosemont–La Petite-Patrie |
| STL | Saint-Laurent |
| SLE | Saint-Léonard |
| VER | Verdun |
| VIM | Ville-Marie |
| VSE | Villeray–Saint-Michel–Parc-Extension |
Watch out: The territoire field in dataset metadata uses these short codes. The ARROND_NOM field in data records uses the full name. Capitalization and accent usage may vary between datasets.
Dash types matter: Some datasets use en-dash (–) between compound borough names, others use hyphen (-). When filtering, try both or use LIKE '%Mercier%Hochelaga%'.
Pattern 4: Output Formatting / Formatage de sortie
Detecting user language
Determine the user's language from their query. If ambiguous, default to the language they wrote in.
Presenting results bilingually
When bilingual fields are available:
🌳 Silver Maple / Érable argenté
Diameter: 45cm | Borough: Ville-Marie
When only French data available, present it naturally:
📍 Location: Banquette gazonnée (grassy median)
Street: Rue Sainte-Catherine
Units and conventions
- Distances: metric (km, m) — Montréal uses metric
- Coordinates: WGS84 (latitude, longitude) preferred; some datasets use NAD 83 MTM Zone 8
- Dates: ISO 8601 (
2024-01-15) or Montréal convention (15 janvier 2024 / January 15, 2024)
- Currency: Canadian dollars (CAD / $)
Pattern 5: API Metadata Language / Langue des métadonnées API
Dataset metadata
title — French (always)notes — French (always), often includes detailed data dictionarytags — French (always)resources[].name — French (always)resources[].description — French (always)groups[].title — French (always)organization.title — French (proper name)
The portal itself
donnees.montreal.ca — French interfacedonnees.montreal.ca/en — English interface- API responses are language-independent (same JSON regardless of language)
Practical implication
When using package_search, always compose queries in French for best results. If the user asks in English, translate the search terms to French, run the query, then present results in English.
Common Translation Table: Dataset Field Names
These are recurring field names across multiple datasets:
| French Field | English Meaning | Notes |
|---|
| NOM | Name | Often a proper noun |
| ARROND / ARROND_NOM | Borough / Borough name | |
| Rue | Street | |
| No_civique | Civic number (house #) | |
| Latitude / Longitude | Same in both languages | WGS84 |
| Coord_X / Coord_Y | X/Y Coordinates | NAD 83 MTM8 |
| Date_debut / Date_fin | Start date / End date | |
| Date_Releve | Survey date | |
| Date_Plantation | Planting date | |
| Statut | Status | |
| Type | Type | |
| Description | Description | Usually in French |
| Adresse | Address | |
| Code_postal | Postal code | |
| Superficie | Area (m²) | |
| Population | Population | |
Provenance / Provenance
| Field | Value |
|---|
| Publisher | N/A — this is an agent-side skill, not a data source |
| Government level | Applies across all tiers (municipal, provincial, federal) |
| Jurisdiction | Montréal / Québec |
| Legal context | Québec's Charter of the French Language (Bill 101) makes French the official language. All government data is published in French first; English availability varies. |
| Last verified | March 2026 |
Accountability note: When presenting translated field names or values, the agent should indicate when it is providing its own translation vs. using an official bilingual field (e.g., Essence_fr / Essence_ang). Never present an agent translation as official government content.
Related Skills / Compétences connexes
understand-ckan — API conventionsdiscover-datasets — Search in the right languageborough-context — Full borough reference with boundarieserror-recovery — Handle encoding and accent issues
