OvercookedChat : A Japanese Human-Human Dialogue Dataset from a Real-Time Collaborative Cooking Game.

OvercookedChat is a human-collected Japanese dialogue dataset for real-time collaborative tasks. This dataset was collected through collaborative work between players with pre-assessed task proficiency in a real-time cooperative cooking game environment inspired by the popular game "Overcooked." This dataset was collected using an offline cooperative cooking game environment specifically designed to support controlled two-player experiments.
➡️ The implementation of the data collection environment is available here:
🔗 https://github.com/UEC-InabaLab/OvercookedChatEnviroment

This work has been accepted to the main conference of SIGDAIL 2025.

🌟 Key Features

• 111 Game Dialogues : A collection of dialogues from 54 participants engaged in a real-time, high-pressure collaborative cooking game.

• 3,412 Annotated Utterances: Captures spontaneous, spoken Japanese communication, with each utterance tagged with one of 16 dialogue acts (e.g., Inform, Request, Suggest).

• Rich Multi-Modal Data: Synchronizes dialogue with 48,781 in-game actions, game scores, and detailed post-session feedback for a holistic view of the collaboration.

• Post-Session Subjective Feedback: Gathers player perspectives on cooperation, communication, and adaptation through an 8-item, 5-point Likert-scale questionnaire after each game.

📊 Dataset Statistics

Category	Total
# Dialogues (Games)	111
# Speakers (Participants)	54
# Utterances	3,412
# Words	22,844
Avg. utterances/dialogue	30.74
Avg. words/utterance	6.69
# In-Game Actions	48,781

📄 Citation

If you use this dataset, please cite the following paper:

@inproceedings{overcookedchat,
  title     = {Task Proficiency-Aware Dialogue Analysis
 in a Real-Time Cooking Game Environment},
  author    = {Kaito Nakae and Michimasa Inaba},
  booktitle = {Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue},
  year      = {2025},
  url       = {https://github.com/UEC-InabaLab/OvercookedChat}
}

📁 Dataset Structure

Each game session is provided as a single JSON file, which contains four primary keys:

• log: A chronological array of timestamped game actions and dialogue utterances, including player IDs and dialogue act tags.

• score: The final game score, including successful/failed dishes and a normalized Z-score.

• survey: Post-session survey results with 8 Likert-scale questions on collaboration and free-form feedback.

• players: Static participant data, including pre-measured proficiency, gaming experience, and personality traits.

⚖️ License

Overcookedchat is released under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license.