DEPRESS: Dataset on Emotions, Performance, Responses, Environment, and Satisfaction during COVID-19.

The DEPRESS dataset covers a 12-month longitudinal period during the COVID-19 pandemic involving 184 undergraduate students.

• Data collection spanned from June 2020 to June 2021.
• Participants were undergraduate students at Worcester Polytechnic Institute.
• Sample size was 184 undergraduate students.
• The study targeted college students as a population particularly vulnerable to pandemic-related mental health disruption.

The dataset is multimodal, incorporating data from online surveys, video recordings, IoT indoor environmental sensors, and Fitbit wristbands.

• Data modalities include demographic and socioeconomic status information, mental health outcome measures, online student engagement metrics, computer and Internet performance data, and daily activity diaries.
• Additional modalities include general indoor environment satisfaction, Fitbit physiological data, sensor-measured indoor environment quality, facial expression data, and GPA.
• IoT sensors measured indoor environmental quality.
• Fitbit wristbands provided wearable sensor data.

The dataset captures mental health outcomes alongside academic performance metrics including GPA during the COVID-19 pandemic.

• Mental health outcome measures were included as a core component of the dataset.
• GPA was collected as an indicator of academic performance.
• Online student engagement was measured to assess the impact of the abrupt transition from in-person to online learning.
• The pandemic severely disrupted students' daily routines due to protective measures and lockdown policies.

To the authors' knowledge, DEPRESS is the first dataset providing multimodal assessment of mental health outcomes, online learning, and potential influencing variables during COVID-19.

• The authors state: 'To our best knowledge, this dataset is also the first dataset that covers multimodal assessment of mental health outcomes, online learning, and potential influencing variables during COVID-19.'
• The dataset integrates environmental, physiological, behavioral, and academic variables simultaneously.
• The inclusion of both objective sensor data and subjective survey data distinguishes it from prior efforts.
• Indoor environment quality was measured by IoT sensors in addition to self-reported satisfaction.

Indoor environment quality and satisfaction were assessed as potential influencing variables on student mental health and performance.

• Both sensor-measured indoor environment quality (objective) and general indoor environment satisfaction (subjective) were collected.
• IoT indoor environmental sensors were deployed to gather objective environmental data.
• Daily activity diaries were also collected to capture behavioral patterns.
• These variables were included to assess the broader environmental context of remote learning during lockdown.

Facial expression data were collected as part of the multimodal mental health assessment during the COVID-19 pandemic.

• Facial expression data were gathered through video recordings.
• Facial expression was included alongside other mental health and behavioral measures.
• Video recordings were one of four primary data collection methods used in the study.
• This modality contributes to the dataset's capacity for non-self-report assessment of emotional states.