Walking along Commercial Drive on a grey Vancouver afternoon, I notice something that’s become increasingly familiar in our cityscape: five fast food chains within a two-block radius, but the nearest grocery store with fresh produce is a 20-minute walk away. For many residents in this pocket of East Vancouver—particularly seniors and those with mobility challenges—accessing nutritious food isn’t just inconvenient; it’s nearly impossible.
A groundbreaking study from MIT’s Computer Science and Artificial Intelligence Laboratory might help explain why such neighborhoods often have higher obesity rates. The research team used artificial intelligence to analyze millions of restaurant menus across the United States, creating what they call a “nutrition landscape” that reveals troubling patterns potentially applicable to Canadian cities as well.
“We’ve known about food deserts for decades,” explains Dr. Marisa Chen, a public health nutritionist at the University of British Columbia who wasn’t involved in the study. “But this research introduces the concept of ‘nutrient deserts’—areas where available food options are uniformly high in calories, sodium, and saturated fat, with minimal nutritional value.”
The MIT researchers developed an AI system that analyzed 14 million dishes from 1.5 million restaurants. By mapping these findings against obesity rates, they discovered areas where residents face systematically limited access to nutritious food options. Neighborhoods with primarily high-calorie restaurant options showed obesity rates up to 15% higher than areas with diverse nutritional offerings.
While conducted in the U.S., these findings likely mirror Canadian reality. According to Statistics Canada, 26.8% of Canadians 18 and older live with obesity, with rates consistently higher in lower-income neighborhoods.
“What makes this study revolutionary is using AI to quantify something public health officials have struggled to measure precisely,” says Dr. Chen. “It’s not just about physical access to grocery stores, but the entire nutrition environment people navigate daily.”
In Vancouver’s Downtown Eastside, I meet Michelle Crow, a community health worker who sees these impacts firsthand. “Many residents here work multiple jobs. Cooking from scratch isn’t realistic when you’re exhausted and your neighborhood only offers fast food that’s engineered to hit all your pleasure receptors,” she explains while helping at a community kitchen.
The situation mirrors problems identified in the MIT study, where researchers found that even when people have theoretical access to healthier options, other barriers remain significant—like cost, convenience, and what they call “familiarity bias.”
“People naturally gravitate toward familiar foods,” notes Dr. James Parker, a food systems researcher at McGill University. “When your neighborhood is saturated with ultra-processed options for years, those become your comfort foods, your defaults. It’s not simply about personal choice.”
This reality challenges the individual responsibility narrative often surrounding obesity. The Health Canada-sponsored Canadian Community Health Survey consistently shows obesity rates following socioeconomic gradients, with lower-income neighborhoods facing higher rates regardless of individual efforts.
In Edmonton, urban planner Sylvia Moreno has been working to address these issues through zoning changes. “We’ve created incentives for grocery stores in underserved areas and limitations on fast-food density,” she tells me during a video call. “But this MIT study gives us a new lens—we need to consider the actual nutritional offerings, not just the type of establishment.”
The AI methodology offers promising applications for Canadian cities. Vancouver’s chief medical health officer recently expressed interest in conducting similar analyses to guide municipal food policy.
The research also highlights how nutrition landscapes vary drastically between neighborhoods mere kilometers apart. In Toronto, for instance, a 2019 University of Toronto study found that residents in high-income neighborhoods had nearly three times the access to fresh produce within walking distance compared to those in lower-income areas.
“This isn’t just about health,” emphasizes Dr. Parker. “It’s about equity. When your postal code dramatically influences your access to nutrition, we’re talking about structural determinants of health that go far beyond personal choice.”
Back on Commercial Drive, I watch a mother with two young children exit a fast-food restaurant. When I later speak with her—Jana, a single parent working two part-time jobs—she explains the reality many families face.
“Of course I know the healthier choice,” she says, adjusting her daughter’s jacket. “But by the time I pick them up from after-school care, I have maybe an hour before bedtime routines. The grocery store means another bus ride, then cooking. Fast food means we actually get family time tonight.”
This perspective illuminates the complex reality behind the statistics. The MIT study found that even when accounting for income and education, the surrounding food environment remained significantly linked to obesity rates.
As Canadian cities grow and densify, understanding these dynamics becomes increasingly crucial. The AI approach pioneered by MIT offers a technological solution to a human problem—using data to visualize invisible barriers that shape our health in ways we often don’t perceive.
“The promise here isn’t just in identifying the problem,” says Dr. Chen. “It’s creating targeted, evidence-based interventions that meet communities where they are.”
Whether this means subsidizing healthy food options, creating new community-based food hubs, or reconsidering urban planning priorities, the goal remains the same: ensuring every Canadian can access nutritious food regardless of where they live.
For now, as I finish my walk through East Vancouver, the disparity remains visible in storefronts and on street corners—a nutrition landscape that, thanks to new research methods, we may finally have the tools to transform.
