News / 17 March 2019

Measuring the mood of the room

Using Twitter to measure the mood of anything is normally pretty dangerous. But it turns out that visitors to Digifest, the leading UK education technology event, are a pretty positive bunch.

Digifest, which takes place every year, helps educators discover the tech trends and facilities that will help students thrive over the coming years. We’d built content and registration platforms for previous years’ events but this time organisers, Jisc, asked us for an immersive installation. They were creating a VR experience and wanted the room that would house the headsets to be an immersive experience in its own right. And they asked for it to demonstrate and create a talking point of AI.

Our solution was to build a multi-sensory sentiment analysis of tweets discussing AI.

We worked with Jisc and their set designers to design the room, adding addressable lighting, a sound system and a series of screens. We then built software based on Stanford’s Natural Language Parser, which scraped Twitter for any tweet containing the hashtag #AI and identified their sentiment.

The tweets, their most meaningful word and their overarching sentiment were shown on the screens. Using colour theory, we then changed the lighting to match the sentiment. And finally, we changed the specially commissioned piece of music that was playing, with positive sentiments triggering sections in the major key, negative sentiments triggering sections in the minor key, and neutral sentiments triggering sections neutral in key.

Jisc loved the installation so much that they have now asked us to install it in the entrance to their new offices, which are due to open later this year.