Datagotchi is an engaging and educational survey tool created by three researchers, designed to help users gain deeper insights into society, especially from an electoral perspective.
With the U.S. elections approaching, Datagotchi is a fun tool that asks questions about your hobbies and interests to predict which political party you’re most likely to support. It’s not all black and white, though—the results are shown as percentages, so you can see how likely you are to vote Democrat or Republican!
The creators wanted to explore the generational shift in voter behavior. Indeed, today, voting patterns are influenced less by social class and religion, and more by lifestyle—what you read, the social media accounts you follow, and even how you dress. In short, they’re shaped by the social groups you identify with.
To learn more about the ideas behind Datagotchi, we spoke with Catherine Ouellet, Assistant Professor in the Political Science Department at Université de Montréal, and one of the creators (along with Yannick Dufresne and Simon Coulombe, Associate Professors at Laval University).
C. Ouellet: We actually started thinking about Datagotchi 7 or 8 years ago. We wanted to show that it’s possible to predict someone’s vote without asking direct political questions. The idea is that we’re all somewhat predictable, and lifestyle choices are markers of socialization that reflect which social groups we belong to—groups that can reveal deeper political tendencies.
C. Ouellet: The U.S. is the ideal setting for Datagotchi. The political and social landscape is extremely polarized, and these divisions often play out in cultural terms—what we call “culture wars.” With strong partisan loyalties, it’s easier to predict voting behavior compared to Canada, where electoral volatility is higher.
C. Ouellet: Yes, we’re adapting the tool for Japan, but it’s challenging because the cultural and political context is so different. We’re collaborating with Japanese political scientists to fine-tune the questions to fit Japanese lifestyles. For instance, in the States, we ask about hunting and fishing, but these wouldn’t resonate in Tokyo. Also, since Japan has been ruled mainly by the Liberal Democratic Party since 1955, predicting vote choice isn’t as relevant. Instead, we’re looking at predicting other attitudes, like the likelihood of voting. Yannick just finished a sabbatical at the University of Tokyo, where he worked with local experts on these challenges.
Datagotchi’s algorithm is powered by data collected via traditional surveys, done by Leger. Catherine and Yannick use this data to train the machine learning algorithm, which then predicts voting behavior based on new user inputs. And we’re constantly improving—after you use Datagotchi, we’ll invite you to validate or correct your prediction. The more people use it, the smarter the algorithm gets!
