mahalisyarifuddin/goemotions-ekman

GoEmotions: Clean Ekman Balanced Dataset (8:1:1 Split)

A clean, single-label, perfectly balanced redistribution of GoEmotions (raw config), mapped to the 8 universal Ekman emotion categories (enjoyment, surprise, fear, sadness, disgust, anger, contempt, neutral).

Standard GoEmotions suffers from extreme class imbalance and synthetic masking tokens ([NAME], [RELIGION]). This redistribution removes placeholder tags, resolves multi-label ambiguity, and provides exact class balance across an 8:1:1 Train/Validation/Test split.

---

Motivation & Applications

Curated for downstream tasks like lightweight text classification, emotional text-to-speech (TTS), and expressive AI behavior.

Modern AI emotion representation faces a strict trade-off:

• Too Limited: 3-class sentiment (Positive/Negative/Neutral) lacks human expressiveness.
• Too Granular: 28+ text labels or infinite paralinguistic TTS speech tags cause severe data sparsity, annotation noise, and control instability.

By standardizing expressive text into 8 universal basic emotions with exact class balance, this dataset provides a robust middle ground for global scalability without heavy computational overhead.

---

Pipeline & Column Schema

1. Tag Filtering: Removed 29,664 comments containing [NAME] or [RELIGION] masking tokens -> 181,561 rows.
2. Ekman Mapping: Collapsed 28 fine-grained emotions into 8 basic Ekman categories following Demszky et al. (2020). Following close investigation, annoyance and disapproval were mapped to contempt.
3. Single-Label Extraction: Retained rows with exactly 1 active Ekman category (ekman_count == 1) -> 161,841 rows.
4. Balanced 8:1:1 Sampling: Bottleneck class is Disgust (2,529). Sampled exactly 2,520 rows per class (2,520 * 8 = 20,160 total), partitioned into 2,016 Train, 252 Validation, and 252 Test per class.
5. Separate Column Schema: Preserves the exact 37 original columns in their original order (text through neutral). Appends separate prefixed indicator columns (ekman_label, ekman_enjoyment, ekman_surprise, ekman_fear, ekman_sadness, ekman_disgust, ekman_anger, ekman_contempt, ekman_neutral) to guarantee zero collision between fine-grained labels and merged Ekman categories.

---

Usage

from datasets import load_dataset

# 1. Load default balanced 8:1:1 dataset (20,160 rows across train/val/test)
ds = load_dataset("mahalisyarifuddin/goemotions-ekman")
print(ds["train"][0]["ekman_label"])

# 2. Or load full unsubsampled single-label pool (161,841 rows)
ds_full = load_dataset("mahalisyarifuddin/goemotions-ekman", "full")

---

Distributions & Mapping

Balanced Split (balanced config — 20,160 rows)

Every split maintains exact 12.5% class balance:

• Train (train.parquet): 16,128 rows (2,016 / class)
• Validation (validation.parquet): 2,016 rows (252 / class)
• Test (test.parquet): 2,016 rows (252 / class)

Full Single-Label Pool (full config — 161,841 rows)

Unsubsampled clean pool in full/train.parquet: enjoyment (62,231), neutral (46,168), surprise (17,709), contempt (14,807), sadness (11,386), anger (4,437), fear (2,574), disgust (2,529).

Mapping Taxonomy

• enjoyment: amusement, excitement, joy, love, desire, optimism, caring, pride, admiration, gratitude, relief, approval
• surprise: realization, surprise, curiosity, confusion
• fear: fear, nervousness
• sadness: remorse, embarrassment, disappointment, sadness, grief
• disgust: disgust
• anger: anger
• contempt: annoyance, disapproval
• neutral: neutral

---

License & Attribution

Distributed under the Apache License 2.0, preserving Google Research's original license. Per Section 4(b), notice is given of tag filtering ([NAME]/[RELIGION]), 28 -> 8 Ekman aggregation, single-label filtering, and stratified 8:1:1 balanced sampling.

@inproceedings{demszky2020goemotions,
  author = {Demszky, Dorottya and Movshovitz-Attias, Dana and Ko, Jeongwoo and Cowen, Alan and Nemade, Gaurav and Ravi, Sujith},
  booktitle = {58th Annual Meeting of the Association for Computational Linguistics (ACL)},
  title = {{GoEmotions: A Dataset of Fine-Grained Emotions}},
  year = {2020}
}