As part of the "Matchbook One" project, Susanne Imhof wants to work with the FHNW to develop a platform for networking residents in retirement homes. Her aim is to counteract the loneliness felt by many people in such institutions and thus increase their well-being. To lay one of the necessary foundations, the "Matchbook Quest" project developed mini-games to create a social platform profile and implemented simple matchmaking for the profiles created in this way. The focus was placed on the needs and physical capabilities of the target group.
The result is a collection of three different mini-games on different topics. Figure 1 shows the game screen of the labyrinth game. In this mini-game, players move a suitcase to the destination of their choice. When the suitcase touches the walls of the labyrinth, an animation and a sound are played.
Several compatibility features have been implemented to make the games accessible to as many people as possible and to offer people with physical limitations, some of which are age-related, the opportunity to participate. For example, the sound feedback and contrast (see "High Contrast Mode" of the labyrinth game in Figure 2) can be adjusted in the games. In addition, the mini-games were optimised for tablet computers with touch input to make it easier to interact with the game elements despite tremors or weaknesses in hand-eye coordination.
A social platform profile is automatically filled in the background based on the decisions made and preferences noted in the mini-games. Based on the profile details, all profiles are categorised into groups with similar interests using the Louvain algorithm ("community detection").
Technologie
Humhub is used as the social network platform. This is based on the Yii2 framework and runs on an Apache web server. The two modules "MBQGames" and "MBQ-Match" are installed on the Humhub instance. These can view and manipulate the profile data via the Humhub API module. "MBQGames" manipulates the profile data in the database and "MBQ-Match" accesses this data in order to execute the matchmaking algorithm. The "PIXI.JS" rendering engine, which uses JavaScript as the programming language, is used to develop the mini-games. The Louvain algorithm, which categorises the profiles into communities, is used to assign the matches.
The user interacts with the two modules via the web application. When a game has been completed, the information collected by the game is saved in the Humhub database via the API module using an API call. With "MBQ-Match", the API call is made directly when the module is started. A bearer token is used for authorisation.
