What does data tell us about off-street parking in Metro Vancouver?

Although parking usually isn't the first thing that comes to mind when thinking about sustainable transportation, where and how much we park plays a significant role in shaping how our cities move and grow.

We recently partnered with TransLink to do a deep dive on non-residential off-street parking across 11 Metro Vancouver study areas. This is the most comprehensive study we know of so far to look at non-residential parking supply, occupancy, and pricing patterns across the region.

A bivariate map showing the intersection of parking intensity and proximity to transit, for non-residential parcels in the study areas.

Key Findings: Metro Vancouver Off-Street Non-Residential Parking

• Transit proximity matters: Developments closer to SkyTrain, SeaBus, and West Coast Express stations tend to have less parking. This is a clear sign of shifting development practices that prioritize transit-oriented growth.

• Parking trends over time: Newer buildings, particularly those built after 2010, show more varied patterns regarding the provision of parking influenced by zoning reforms and major transit expansions. In areas like Metrotown and along the Evergreen Extension, reduced parking ratios suggest that developers are responding to shifting policy and market signals favoring transit access over car storage.

• Suburban oversupply: Many suburban areas maintain high parking vacancy rates, suggesting room to rethink designing minimum parking requirements for peak utilization and better align supply with real demand. Building on the patterns observed near transit, repurposing underutilized parking areas for new transit infrastructure could further reduce parking demand and make more efficient use of land.

• Machine learning potential: Using high-resolution aerial imagery with advanced deep learning algorithms we were able to detect and classify over 20,000 commercial surface parking stalls across the region. Our analysis identified spatial patterns and helped fill in gaps in data where traditional counts aren’t feasible, and demonstrated the role machine-learning can play progressing data input and analytics in regional-planning.

Read on to learn about how we came to these conclusions.

This map shows the spatial distribution of occupancy types, or use, of non-residential parcels throughout the study areas.

Smart Data for Smart Cities

Parking takes up a surprising share of urban space, but its use is often poorly understood. In light of this lack of understanding, we aimed to create information out of data that would help plan for growth that aligns with transit access and land use priorities.

Study areas showing orthoimagery used for machine learning-based detection of parking space

Our team combined machine learning with municipal data and GIS analysis to analyze over 20,000 parking spaces across 11 study areas. Our approach reveals significant differences in parking behaviour and availability across commercial, institutional, and industrial sites across increasing proximity to transit.

In some suburban areas further from transit, large portions of parking supply are under-used, whereas sites closer to SkyTrain stations show tighter, more efficient use of space. These contrasts highlight how parking reform could unlock valuable land for housing, public space, or alternative modes of travel. By grounding this analysis in measured data rather than assumptions, we can help our clients and their partners modernize their parking policies and align them with emerging transit and climate priorities.

The real value of this work isn’t just what it shows, it’s what it enables. By connecting spatial data, machine learning, and existing policy, we’re helping to bridge the gap between technical analysis and planning outcomes. As cities across the region adapt to growth and changing travel patterns, tools like this offer a clearer view of how land is used, and a benchmark for trends moving forward. For us, that’s what “smart data” really means: turning complex spatial patterns into insights that make city-building a little bit smarter.

Building area per parking stall (m²) of buildings constructed after 1990 and distance from transit stations. Each subplot shows a different study area. The horizontal axis indicates distance from the nearest transit station, and the vertical axis shows the mean building area per parking stall (in square meters). A lower building area per parking stall suggests more parking availability relative to the building on the same property.

What’s Next

Our findings point to a broader opportunity: using data-driven insights to guide parking policy reform and land use planning across the region. Updating parking minimums in transit-rich areas, exploring conversions of underused lots, and benchmarking against global best practices could all help Metro Vancouver advance its climate and mobility goals.

Ultimately, this work provides a foundation for a more strategic and sustainable approach to parking management, supporting vibrant, connected, and efficient communities.