The Emotional Internet

Technology is de-humanising the way we lead our lives, therefore it may appear surprising that there is a project seeking to mine emotional data from […]

Technology is de-humanising the way we lead our lives, therefore it may appear surprising that there is a project seeking to mine emotional data from the internet. However, instead of writing in journals or collecting personal family albums, people are sharing an ever-increasing depth of feeling by writing blogs, posting social network statuses and by sharing photos online. Consider that each upload may be accompanied by information regarding time and geographical location, and the web holds a Goliath archive of human communication and emotion. Many sentiment data-extraction and -visualisation projects have emerged which innovatively harvest and present the data in meaningful and beautiful ways.

One such project is called We Feel Fine, the brainchild of a computational mathematician, Sepandar Kamvar, and computer scientist/anthropologist/artist Jonathan Harris. The system trawls through webpages and extracts pages with occurrences of the phrases “I feel” and “I am feeling”. It then identifies the ‘feeling’ in the sentence (e.g. excited, sad etc.) and mines this along with the age, gender and location of the author which can be extracted relatively easily from blogs. The system extracts around 10,000-15,000 new feelings each day, adding to a database that already contains several million emotions. If there are images in the post, the indexer extracts the largest image and sends it to an image repository. Where location information is available, the indexer sends the time and geographical information to a weather server that acquires the weather using several public weather databases.

The full experience of the site can be viewed online, and Harris presents some of the main elements in his Ted talk. A paper has also been published. The documented response of users to interacting with the system is that they feel a heightened self-awareness and self-reflection of their own sentiments, as well as a connection to the authors of the emotional data they are viewing. There are also some curious emotional trends that emerge, for example that fact that people become considerably less angry, disgusted and sad as they get older. Of course, the primary limitations of these systems are that they are exclusive of people or emotions that are not documented online, they don’t address the question of whether people are genuine in their online expressions, and there are ethical issues around whether or not the emotional data posted on the web should be used in these compositions. Nevertheless, the ethnographic motivation for these types of projects cannot be ignored, and the results are stunning. Swimming in the same stream, emoto is an online service created through collaboration between designers and engineers which presents interesting visual compositions that document the emotional response of online users to London 2012.

These types of projects allow us to answer those questions you can’t Google such as: “What was the overriding sentiment in the UK on the third day of the Olympics 2012?”  and in a small way help us to be more connected to our emotions and to society.

About Anna Zawilska

I am reading for a DPhil in Computer Science concentrating on technology and education. My research interests lie around the intersection between technology and social issues.