Sandrine Müller uses smartphone sensor data to study human behavior.

A research team led by Sandrine Müller, a former Data Science Institute (DSI) postdoctoral research fellow, found that depression predictions from GPS-based mobility do not generalize well to large demographically heterogeneous samples.

Müller, who recently joined Google as a quantitative UX researcher, and collaborators, including Xi (Leslie) Chen, a doctoral student in computer science; Heinrich Peters, a Columbia Business School (CBS) doctoral candidate; Augustin Chaintreau, an associate professor of computer science and DSI member; and Sandra C. Matz, David W. Zalaznick Associate Professor of Business at CBS and DSI affiliate, published their findings in Scientific Reports. 

READ THE FULL ARTICLE

Depression is one of the most common mental health issues in the U.S., affecting the lives of millions of people suffering from it as well as those close to them. Recent advances in research on mobile sensing technologies and machine learning suggest that a person’s depression may be passively measured by observing patterns in people’s mobility behaviors. However, the majority of work in this area has relied on highly homogeneous samples, most frequently college students.

In this study, Müller, et al. analyzed more than 57 million GPS data points to show that the same procedure that leads to high prediction accuracy in a homogeneous student sample, leads to accuracies only slightly higher than chance in a U.S.-wide sample that is heterogeneous in its socio-demographic composition as well as mobility patterns. This pattern holds across three different modelling approaches which consider both linear and non-linear relationships. Further analyses suggest that the prediction accuracy is low across different socio-demographic groups, and that training the models on more homogeneous subsamples does not substantially improve prediction accuracy. Overall, the findings highlight the challenge of applying mobility-based predictions of depression at scale.

ABOUT THE DATA SCIENCE INSTITUTE AT COLUMBIA UNIVERSITY
The Data Science Institute (DSI) at Columbia University was established in 2012 to advance the state-of-the-art in data science; transform all fields, professions, and sectors through the application of data science; and ensure the responsible use of data to benefit society. We train the next generation of data scientists, develop innovative technology, foster collaborations in advancing techniques to interpret data and address pressing societal problems, and work closely with industry to bring promising ideas to market. Discover more at datascience.columbia.edu.