Does spending time in beautiful settings boost people’s happiness? The answer to this question has long remained elusive due to a paucity of large-scale data on environmental aesthetics and individual happiness. Here, we draw on two novel datasets: first, individual happiness data from the smartphone app, Mappiness, and second, crowdsourced ratings of the “scenicness” of photographs taken across England, from the online game Scenic-Or-Not. We find that individuals are happier in more scenic locations, even when controlling for a range of factors such as the activity the individual is engaged in at the time, weather conditions and the income of local inhabitants.

However, what might these beautiful places be comprised of? Is beauty in this context synonymous with nature? We extract hundreds of image features from over 200,000 Scenic-Or-Not images using the Places Convolutional Neural Network to understand the composition of beautiful places. We also find that a neural network can be trained to automatically identify scenic places, including both natural and built locations.

Does spending time in beautiful settings boost people’s happiness? The answer to this question has long remained elusive due to a paucity of large-scale data on environmental aesthetics and individual happiness. Here, we draw on two novel datasets: first, individual happiness data from the smartphone app, Mappiness, and second, crowdsourced ratings of the “scenicness” of photographs taken across England, from the online game Scenic-Or-Not. We find that individuals are happier in more scenic locations, even when controlling for a range of factors such as the activity the individual is engaged in at the time, weather conditions and the income of local inhabitants.

However, what might these beautiful places be comprised of? Is beauty in this context synonymous with nature? We extract hundreds of image features from over 200,000 Scenic-Or-Not images using the Places Convolutional Neural Network to understand the composition of beautiful places. We also find that a neural network can be trained to automatically identify scenic places, including both natural and built locations.

Does spending time in beautiful settings boost people’s happiness? The answer to this question has long remained elusive due to a paucity of large-scale data on environmental aesthetics and individual happiness. Here, we draw on two novel datasets: first, individual happiness data from the smartphone app, Mappiness, and second, crowdsourced ratings of the “scenicness” of photographs taken across England, from the online game Scenic-Or-Not. We find that individuals are happier in more scenic locations, even when controlling for a range of factors such as the activity the individual is engaged in at the time, weather conditions and the income of local inhabitants.

However, what might these beautiful places be comprised of? Is beauty in this context synonymous with nature? We extract hundreds of image features from over 200,000 Scenic-Or-Not images using the Places Convolutional Neural Network to understand the composition of beautiful places. We also find that a neural network can be trained to automatically identify scenic places, including both natural and built locations.