Share And Promote Your Article


Google Maps used to detect factors that favor obesity in cities

Nutrition and Diet

Author marstighter Page Views 109 3 min read

Google Maps used to detect factors that favor obesity in cities

They have used a neural network of artificial intelligence to relate the prevalence of obesity in some cities in the United States with different environmental factors, such as the presence of gyms, parks or green areas.

In broad strokes, we look for infrastructures that make our lives easier so that we have to move as little as possible during our day today. That is clear, but is it the only thing to pay attention to?

According to a research that is published in the JAMA Network Open, by the hand of scientists from the University of Washington, other factors should be considered, at least if you want to find a place where healthy life prevails. They have demonstrated this thanks to a model of artificial intelligence capable of relating the prevalence of obesity within the built environment.

 

Detect obesity from space

In 2016, there were more than 340 million children and adolescents overweight or obese worldwide, according to data from the World Health Organization. Since 1975, obesity has almost tripled worldwide. The global obesity levels have nearly tripled since 1975 and are advancing by leaps and bounds, fueled by factors derived from the current lifestyles, such as sedentary lifestyle or increasing the supply of fast food.

All specialists agree that obesity is influenced by apparent factors, such as genetics, diet, physical activity, and the environment. However, evidence pointing to associations with the built environment has varied widely between studies and geographical contexts. Therefore, these researchers have decided to seek that association, with the help of Google Maps and artificial intelligence.

google map

Relation of google map and artificial intelligence

The study is based on the use of a convolutional neuronal network; that is, a type of artificial intelligence that exploits deep learning to identify particular patterns in a dataset independently.

In this case, the objective of this model was to locate patterns associated with obesity in images of American cities taken via satellite with the popular Google application. In this way, the neural network learned to relate to both factors, being able to predict later the association. The model analyzed factors such as the presence in the area of gyms, spas, bakeries, supermarkets or bowling alleys.

The model analyzed specific points of the built environment that could be related to the prevalence of obesity, such as the presence of gyms or bowling alleys. It also focused on the existence or not of green areas and nearby parks, as this could encourage people who live in the area to exercise outdoors.

Finally, all these characteristics of the built environment accounted for 64.8% of the variation in the prevalence of obesity in the districts analyzed.

 

Other data of interest

The model also found some linkage with the socioeconomic level of each district. Specifically, in areas with a higher figure of income per capita prevalence estimates of age-adjusted obesity they were lower.

This is because a person whose income is barely enough to live or spend money every month in a gym, no matter how close he lives. Either way, it would not stop you from running or doing some other type of exercise at home. The next step of these scientists will be to focus on the influence of this type of factors, more deeply.

It is clear that people who live in green and open areas are more likely to exercise, while those who live in densely populated areas without vegetation often lead a more sedentary and busy life, which prevents them from exercising unless they pay for it. 

All this could be intuited. However, with this model, the most problematic areas can be detected, to accurately direct a higher number of information and prevention measures to them. It seems that Google Maps is a tool that has many applications to explore.