That place that makes you happy? New Technion algorithm can pinpoint it
One person may have “left his heart in San Francisco” or think that “London is waiting for him,” while another calls Jerusalem the “City of Gold.”
Now, scientists at the Technion-Israel Institute of Technology’s Faculty Civil and Environmental Engineering and AT&T’s Research Laboratories have found the connection between our geographical location and the emotions that arise in us – which places tend to make people happy and what locations make them angry or sad.
The study, based on communications in the social networks such as Twitter, was conducted as part of the doctoral work of research student Ben Galon, led by Prof. Yerah Deutscher of the Technion and Dr. Yaron Kanza of AT&T.
Until now, emotional maps have been based on monitoring volunteers by means of sensors that measure heart rate, blood pressure and other physical factors. This method severely limits the amount of information that can be collected. Now, thanks to the new Technion approach, you can create maps based on vast amounts of information.
The researchers have succeeded in linking emotion to places by means of a syntactic analysis of messages containing location tags. The algorithm they developed enables us to locate significant connections between emotion and place, and to map out the random connections.
The map of emotions produced by the technology presents the characteristic emotions expressed in different places. Such maps offer a variety of possible uses; for example, they can help a tourist choose where he or she is likely to feel joy or romantic feelings and avoid places that induce anger and hostility.
City planners will be able to identify areas where planning triggers negative emotions, and change them to induce other emotions. Social scientists will be able to learn about relationships between community behavior and their environment, and identify areas that intensify artistic feelings or anger that can lead to political activism.
The study dealt with three main challenges. First, social media messages are often short, poorly written and contain expressions, abbreviations and slang, making it difficult to accurately characterize them. Second, a wide range of emotions is expressed everywhere, so it is important to filter out the meager and random bonds and leave only the significant and significant correlations. Third, to get meaningful results in a short time, a speedy algorithm is needed that can handle huge amounts of data.
The proposed method was tested on a database of tens of millions of messages created in New York City. Anger, for example, was detected at a high level in public transport stations; sadness and anger were found in high schools. Anxiety was identified on university campuses but also in Madame Tussaud’s wax museum. Joy, sometimes accompanied by surprise, was linked to restaurants and parks. All the emotions examined were very frequent in theaters and cinemas.
According to Galon, the study focused on an in-depth analysis of the results to understand the reasons for the links found between emotion and place and to introduce the time dimension into the new emotion maps. The study was conducted in the US in part because of the great use made by Americans of networks such as Twitter, but the technology is relevant wherever these networks are widely used, he said.