BUSINESS INSIGHTS : The latest news, analysis, and trends about protection and care

Sep 27, 2018,  by Allianz Partners Business Insights

AI can estimate the obesity rate in a geographical area thanks to Google Maps

By analysing the information gleaned from Google Maps and Street View, as well as data from previous studies, an algorithm is able to determine the obesity rate in a particular geographical area according to its urban architecture. In this way, artificial intelligence has uncovered new determining factors when it comes to obesity.

There is now an artificial intelligence (AI) that is capable of determining the obesity rate of a neighbourhood’s population… based on data drawn from Google Maps. The algorithm was developed by two researchers from the University of Washington (United States). They based their study on the relationship between the prevalence of obesity in a community and its social and economic environment, as well as its urban features.


Data concerning six large American cities


Ayasha Maharana and Elaine Okanyene Nsoesie carried out their research on six large American cities, reported a scientific article published on 31st August in the Journal of the American Medical Association (JAMA). The AI was provided with information about the structure of the cities concerned and data about the causes of obesity drawn from other studies. The algorithm was then put to work and managed to identify the areas with the highest prevalence of obesity by analysing the mass of data supplied. 

150,000 images from Google Maps and Street View were entered into the automated learning system, as well as several million “points of interest”. These included sports facilities, pet shops and parks, which enabled the AI to refine its estimations.


Using the number of buildings and type of accommodation as obesity indicators


Thanks to the artificial convolutional neural network (CNN) used by the researchers, the programme highlighted a correlation between the material wealth of a community and a low prevalence of obesity, a trend that had already been recognised, reports L’ADN. But the artificial intelligence showed that there were other determining factors at play, such as the number of facilities dedicated to health or wellbeing in the neighbourhood. The number of buildings also affects the results, as does the predominant type of residence.

The American researchers’ study also showed that some disadvantaged geographical areas had a lower obesity rate thanks to their built environment. In the future, the tool may be used to establish new ways of fighting obesity.


Allianz Partners

Linked Topics
About Allianz Partners Business Insights