DiversityOne: A Multi-Country Smartphone Sensor Dataset for Everyday Life Behavior Modeling

As part of its ongoing commitment to advancing diversity-aware, person-centric, and data-driven artificial intelligence, the DataScientia Initiative is pleased to announce an upcoming lecture by Prof. Fausto Giunchiglia on the DiversityOne dataset.

On 11 September 2026, Prof. Giunchiglia will present “DiversityOne: A Multi-Country Smartphone Sensor Dataset for Everyday Life Behavior Modeling” at the Forum of Great Minds at the University of Science and Technology of China (USTC).

Understanding Everyday Life Across Countries

Understanding people’s everyday behaviour is important for a wide range of AI applications. However, existing datasets are often limited to individual countries, relatively small participant groups, or a narrow selection of pre-processed sensor data. These limitations make it difficult to study how human behaviour varies across countries and to evaluate the generalization and robustness of AI models.

DiversityOne addresses this gap by providing a large, multi-country dataset covering eight countries: China, Denmark, India, Italy, Mexico, Mongolia, Paraguay, and the United Kingdom.

The dataset includes data collected from more than 18,000 participants, including over 350,000 self-reports from 782 participants across eight countries, together with data from 26 sensor types. It combines smartphone sensor information with extensive demographic and psychosocial survey data, offering a diverse foundation for research into everyday human behaviour.

During the lecture, Prof. Giunchiglia will discuss the research that led to the creation of DiversityOne, highlight selected research results enabled by the dataset, and outline opportunities for future research. The talk will also demonstrate how cross-country and culturally diverse datasets can contribute to the development of AI systems that better account for differences in human behaviour, experiences, and contexts.