Why City Planners Might Want to Talk to More Women


Does danger really lurk in darkened alleyways and on desolate streets? Most of us just trust our gut when wandering in unfamiliar areas, but researchers are using a machine learning-enhanced image database and cell phone data to determine urban characteristics that people seem to fear most, which they believe could help with city planning.

In a paper presented last week at the Association for Computing Machinery's Multimedia Conference, researchers at the MIT Lab, as well as the University of Trento and the Bruno Kessler Foundation in Italy, explained that "safety scores" assigned by people—and now software programs—to images of Rome and Milan in a database lined up nicely with areas that people did or did not frequent regularly, as determined by cell phone tower data.

For example, buildings with street-facing windows created "a sense of safety, since they imply the possibility of surveillance," they found. Well-kept green spaces were also perceived as safe, though poorly maintained green spaces were not. In the image below, the photo on the left (a darkened tunnel) was determined to be most unsafe, while the one on the right (a bright office park) was deemed most safe.

Based on cell phone tower data provided by Telecom Italia Mobile—adjusted for factors such as population density and distance from city centers—the areas rated as looking unsafe were visited less frequently. Those visitation rates were strongest among women and people over 50 years of age because men under the age of 30 were more likely to visit unsafe neighborhoods, according to the paper.

Luis Valenzuela, an urban planner and professor of design at Universidad Adolfo Ibáñez in Santiago, Chile, tells MIT News that the "research gives scientific evidence that women have a specific perception of the appearance of safety in the city," which city planners might want to take into consideration.

For years, images in the database were voted on by volunteers, but three years ago, researchers started using the human responses to train a machine-learning system that would assign scores.

"That's ultimately how you're ble to take this type of research to scale," César A. Hidalgo, an assistant professor at the MIT Media Lab, said in a statement. "You can never scale by crowdsourcing, simply because you'd have to have all of the Internet clicking on images for you."

Looking ahead, the researchers say more study is required. However, they do believe that as the research grows and more cities are evaluated, the combination of cell phone data and machine learning could improve and beautify cities.

0 Comment

Leave a Reply

Captcha image


  • 5300c769af79e

    Nest Store Selling Compatible Connected Home Products

    No longer will the store just carry Nest's own connected home products.The company this week bolstered its digital store shelves with some third-party "Works with Nest"-approved products, such as LIFX Wi-Fi smart bulbs and the Philips Hue Starter Kit.
  • 5300c769af79e

    Qualcomm Unveils the Full Snapdragon 821 Specs

    After a brief introduction to the Snapdragon 821 back in July that lacked specific details about its advantages over the Snapdragon 820, Qualcomm today revealed it all just as IFA was kicking off in Berlin.According to today’s series of press releases, the Snapdragon 821 (MSM8996Pro) is 10% faster in boot times than the 820, can reduce app load times by 10%, improves general performance by up to 10%, and provides smoother UI experiences and browsing performance.
  • 5300c769af79e

    Identity Crisis: How to Balance Digital Transformation and User Security

    Download Organizations are facing the need for improved capabilities in managing identities and access as part of their journey through Digital Transformation.Managing customer identities becomes the standard, beyond traditional employee-focused Identity and Access Management (IAM).
  • 5300c769af79e

    Order Pizza Hut With This Temporary Tattoo

    Thanks to Pizza, Hut, some people in the UK will soon be able to say yes to that question.The beloved pizza chain has invented what might be the most hipster thing ever: a temporary tattoo that can order pizza.