Accessible online since Tuesday morning, the Datagotchi application was designed by the research team of the Leadership Chair in Digital Social Science Teaching at Laval University (CLESSN).
After comparing the effectiveness of different statistical models, the designers of the web platform, whose name is a nod to Tamagotchi, that virtual pet that was all the rage in the 90s, trained an algorithm to learning capable of predicting users’ voting intentions from a list of 30 questions.
These range from sexual preferences to the most frequently used mode of transportation, hair length, type of dwelling and level of interest in hunting and outdoor sports. A user’s avatar takes shape as they complete the questionnaire.
Revealing variables
If the application uses classic socio-demographic characteristics such as age, gender and place of residence, it is first and foremost focused on the user’s lifestyle, insists Catherine Ouellet, doctoral candidate. in political science at the University of Toronto and collaborator on the project.
This is to show, roughly, that these variables, where we stand, what we do, what we eat, etc., can tell more about who we are and what ‘we think and can be especially rich, when we are interested in politics, then it can also be variables that will be used by strategists, she explains in an interview on the show All morning.
Catherine Ouellet explains that the application aims in particular to make citizens aware of what their lifestyle can reveal about themselves.
The researcher specifies that there is no direct relationship between, for example, the preferences of a voter and his taste for latte. Lifestyle habits are approached in the same way as traditional socio-demographic variables, whose predictive power tends to decline.
It is not the fact of drinking a latte that makes me vote, for example, conservative. It’s not watching this kind of movie that makes me vote for the Green Party, it’s an indicator of something else, insists Ms. Ouellet.
These are markers [ou] units of socialization that can allow us to delimit new groups of the electorate.
The team behind the creation of the Datagotchi app also contributed to the development of the Voter Compass (archive).
In this digital age, where every transaction made on Amazon and every viewing made on Netflix leave traces, the researcher believes it is important for voters to understand that their lifestyles can be reclaimed for political ends.
We really try to make citizens aware of the wealth of this data and then of the information they can give about us, without necessarily always realizing it., insists Catherine Ouellet.
Like the other members of the team behind the creation of Datagotchi, she hopes that the application will enter into the mores of Quebec and Canadian voters, just like the Electoral Compass, to which the CLESSNLeadership Chair in Digital Social Science Education at Laval University had also contributed.
To access the Datagotchi platform, click on this link (New window).