Oncology Support Voice Dataset: Emotionally Balanced Customer Service Calls

**Overview**
This dataset features 50,000 authentically synthesized audio recordings from 500 diverse speakers, each performing 100 scripted oncology-related customer service sentences. The scripts are designed to capture a range of emotions—worried, frustrated, relieved, sad, grateful, and neutral—and are authentically voiced to simulate real-world healthcare support interactions. The dataset includes both high-quality audio and telephone-style filtered versions, making it ideal for testing empathetic AI agents and testing in call center environments.

**Use cases**
- Train virtual agents for emotionally intelligent oncology support and healthcare helplines
- Build expressive TTS systems for caregiving, wellness, or telehealth platforms
- Detect vocal emotional markers for patient support and engagement
- Create datasets for research in human-computer emotional understanding

**Product details**

- Datasets represented include oncology_emotion_calls.csv, oncology_audio_original or phone.
- Sample fields include text, emotion_label, audio_file_path (original or phone)

**Additional Insights**
This dataset was created to support the development of compassionate and context-aware AI for high-empathy environments. Scripts are based on real oncology customer service scenarios and emotions are intentionally balanced across speakers and dialogues. All data is ethically sourced, with full consent from participants. The inclusion of telephone-style recordings enables testing for real-world deployment settings, such as telehealth or customer support call centers.

For more details, please feel free to reach out to us at [e-mail]