AI Algorithm Identifies How City Structure Affects Health

A machine-learning algorithm has been developed to estimate obesity levels in US cities without directly assessing the medical data of inhabitants. The researchers hope their findings can help future cities improve the health and wellbeing of their residents.
Researchers from the University of Washington studied satellite and Google Maps Street View imagery of city infrastructure and building placement, correlating it with obesity rates in individual cities. They also included 'points of interest' such as food and pet shops, which encourage activity within a district. For example, in areas with shops, people are more likely to walk around and socialise compared to less-frequented industrial districts.
Their initial research has found, unsurprisingly, that green urban areas with widely spaced buildings correlated with lower obesity rates, as these features facilitate physical activity. Despite wealthy areas typically including these elements, validation tests demonstrated that income was only one contributing factor to inhabitants' health; a city's infrastructure also affected its obesity rates.
The algorithm has only been applied to US cities so far, but could be rolled out further afield if adapted to account for differences in city planning and lifestyle across other cultures.
Obesity affects almost 40% of US adults (CDC, 2018). Dynamic approaches to health management in cities is a wise move, as less than 20% of the US population live in rural areas (Census Bureau, 2016). The University of Washington's research will be helpful in planning future urban infrastructure and offers a novel solution to concerns over healthcare.
Our recent blog on Norwegian town Lyseparken illustrates how cities of the future can be built with the wellbeing of inhabitants in mind. For more on the future of urban spaces, see our Smart Cities Spotlight Trend.