Are MIT students happier on Fridays? Do midterms prevent students from smiling? Does the weather influence the mood of the campus?
The Media Lab’s Mood Meters, which use four displays across campus to gauge the mood of people passing by, aim to answer these questions. Part of the MIT150 Festival of Arts, Science, and Technology (FAST), the Mood Meters analyze the number and intensity of smiles at MIT through an aggregate of facial expression data received via cameras at the four installations. The first Mood Meter went live on April 6 on the third floor of the Media Lab, and since then three more have been installed in the Infinite Corridor, the Stata Center, and the Student Center.
The masterminds of the project are Javier Hernandez G and M. Ehsan Hoque G, two graduate students in the Media Lab’s Affective Computing Group. Hoque and Hernandez’s research focuses on computing emotion and developing technologies that raise awareness of certain affective (emotional) states. FAST provided the two graduate students with support and the opportunity to showcase different elements of their research by placing their work “in the wild,” Hoque said.
Although those who interact with the exhibit may try to display sad expressions, all emotions besides happy are simply registered as neutral. The inventors said that, while they could have made the Mood Meters recognize other expressions, both agreed that they wanted to “cheer up the whole mood of MIT,” according to Hernandez.
People playing with the meter by intentionally cracking a smile or placing pictures in front of the camera are unlikely to affect the results of the project. Though they could create an algorithm to filter out such readings, both Hernandez and Hoque say they are currently interested in seeing the mood of MIT as people interact with the exhibits and as the MIT community becomes more familiar with the technology. To further encourage interaction, they have made all the data they collect during the project visible — in real time — on their website, http://moodmeter.media.mit.edu.
The success of the Mood Meter is apparent from the crowds of visitors seen collecting around the screens and students stopping in the hall to smile and observe how the Meter fluctuates.
People were not always so enthused about the project, especially due to privacy concerns. Prior to the deployment of the project, Hoque and Hernandez met with Information Services and Technology (IS&T), the Security and Emergency Management Office (SEMO), and Audio Visual Services to explain that the Meter does not record images or any personal information of those passing by. Instead, they said, the installation collects data from a live feed of, according to Hernandez, “just the number of smiles” and “just the number of people.”
Hernandez and Hoque were open about the objective of the project from the beginning and maintained that there never was going to be any image recording. Because of this, they said they were able to win over the concerned parties and obtain the permission needed to install the Mood Meters. Hoque said that with “transparency from the very beginning, people can be very supportive and understanding.”
Although Hernandez and Hoque have received countless suggestions on what they could do next with their Mood Meter, such as one Twitter user’s suggestion to “install this meter immediately in the U.S. Senate,” neither have any future plans for the Mood Meter at the moment; instead, they both said they would prefer to focus on research. Nevertheless, they hope that by engaging the community and addressing the issue of privacy, they have served as a voice for this type of technology, paving the way for future applications.
The idea of assessing emotions with the use of technology is not necessarily new. For example, the Dutch Applications Company has created an application for iPhone and Android called Happy Map that questions users in Belgium about their mood and then generates a map of the mood levels across the country. However, such data can be unreliable due to user’s subjectivities. Technology employed in the Mood Meter instead aims to provide a more accurate and objective method of assessing moods and emotions.