The Social-Emotional Music (SEM) Dataset represents a novel contribution to music information retrieval research, specifically addressing the intersection of music psychology and social-emotional learning (SEL). This curated collection provides music tracks systematically categorized according to three fundamental emotional competencies: Outlook, Problem-Solving, and Empathetic Perspective-Taking.
Unlike traditional music datasets that focus primarily on genre classification or basic emotional valence, the SEM Dataset bridges the gap between low-level acoustic features and high-level psychological constructs. The dataset was developed through interdisciplinary collaboration between musicologists, psychologists, and educational researchers to support empirical investigation of music's role in social-emotional development and therapeutic applications.
The SEM Dataset provides a standardized resource for researchers investigating the relationship between musical structure and psychological impact, with potential applications in music therapy, educational technology, and affective computing.
| Property | Value |
|---|---|
| Total Tracks | 419 |
| Total Duration | 3.5 hours |
| Audio Format | WAV |
| Sample Rate | 44.1 kHz |
| Total Size | 2.5 GB |
| License | CC BY-NC 4.0 |
The three-category framework reflects established constructs in social-emotional learning theory:
| Category | Tracks | Duration | Theoretical Foundation |
|---|---|---|---|
| Outlook | 147 | 1.23 hours | Promotes optimistic cognition through melodic coherence, tonal stability, and harmonic consonance |
| Problem-Solving | 138 | 1.15 hours | Enhances analytical thinking through structural complexity, tension-resolution patterns, and cognitive engagement |
| Empathetic Perspective-Taking | 134 | 1.12 hours | Facilitates perspective-taking through polyphonic textures, voice leading, and interpersonal musical dialogue |
Track selection was guided by established theories in music cognition and educational psychology. Each category represents specific psychological mechanisms:
Outlook: Based on research linking melodic fluency and tonal stability to mood enhancement. Selected tracks demonstrate characteristics associated with stress reduction and positive affect, including:
- Predictable melodic patterns that reduce cognitive load
- Consonant harmonic progressions that create emotional stability
- Moderate tempo ranges conducive to relaxation responses
Problem-Solving: Grounded in theories of musical expectation and cognitive flexibility. These tracks feature structural elements that mirror problem-solving processes:
- Clear presentation of musical "problems" (tensions, dissonances)
- Logical developmental processes leading to resolution
- Balanced complexity that engages without overwhelming cognitive resources
Empathetic Perspective-Taking: Informed by research on musical perspective-taking and social cognition. Selection emphasized:
- Polyphonic textures requiring attention to multiple musical "voices"
- Interactive musical elements suggesting dialogue or conversation
- Emotional narrative arcs that model empathetic responses
The SEM Dataset supports multiple research paradigms:
- Classification Research: Automatic categorization of music by social-emotional impact
- Music Therapy Studies: Evidence-based music selection for therapeutic interventions
- Educational Technology: Intelligent music recommendation for social-emotional learning
- Cognitive Science: Investigation of music-emotion relationships
If you use the SEM Dataset in your research, please cite:
@dataset{sem_music_dataset_2025,
title={SEM Music Dataset: A Dataset for Social-Emotional Music Classification},
author={[Authors]},
year={2025},
publisher={[Publisher]},
version={1.0},
doi={[DOI_NUMBER]}
}This research was supported by [Funding Sources]. We acknowledge the contributions of domain experts in music psychology and education who participated in the annotation process, as well as [Institution] for computational resources and infrastructure support.
The development of this dataset reflects collaborative efforts across disciplines, demonstrating the value of interdisciplinary approaches to complex research questions at the intersection of music, psychology, and technology.
If you have any questions regarding the SEM Music Dataset or related research, please feel free to contact:
📧 Prof. Shanken — shanken@cycu.edu.tw