InnerI/consciousness-hashtag #consciousness

Dataset Card for consciousness-hashtag or #consciousness

Dataset Summary

The consciousness-hashtag dataset, curated by Inner I Network (Gravity powered with macrocosmos.ai), is a dynamic collection of 988 English-language tweets sourced from X (formerly Twitter) that contain the word "consciousness" or the hashtag #consciousness. This dataset encapsulates a global dialogue on awareness, weaving together philosophical inquiries, scientific explorations, spiritual awakenings, and speculative discussions on topics like AI sentience, panpsychism, and the nature of the self. From casual reflections ("That moment of pure consciousness at sunrise #consciousness") to deep dives ("Microtubules as the seat of consciousness? Let's debate. #neuroscience"), these tweets form a living archive of human curiosity about what it means to be. Ideal for NLP researchers, philosophers, cognitive scientists, and AI ethicists, this dataset powers tasks like sentiment analysis on existential themes, topic modeling of spiritual trends, and training models to understand consciousness discourse. Whether you're mapping societal shifts in awareness or fine-tuning LLMs for introspective text generation, #consciousness-hashtag is your gateway to the collective mind.

Supported Tasks

• Sentiment Analysis: Gauge emotional tones in consciousness discussions (e.g., awe, doubt, enlightenment).
• Topic Modeling: Uncover clusters like spirituality vs. neuroscience or AI ethics.
• Text Classification: Label tweets by perspective (e.g., scientific, metaphysical, personal).
• Trend Analysis: Visualize spikes in #consciousness usage tied to events like global meditations or tech releases.
• Generative AI: Fine-tune models to produce consciousness-themed poetry or debates.

Languages

Primarily English (en), with filtering to ensure linguistic consistency for NLP applications. Some tweets may include multilingual hashtags, but core content is English-focused.

Dataset Structure

The dataset is structured as a tabular collection of tweets, with rich metadata for contextual analysis. Each row represents a single tweet, including content, user details, and interaction flags.

Example:

{
  "uri": "https://x.com/user/status/1234567890123456789",
  "label": "",
  "username": "conscious_seeker",
  "text": "In the silence of meditation, consciousness unfolds like a cosmic flower. 🌸 #consciousness #spirituality",
  "tweet_hashtags": ["consciousness", "spirituality"],
  "url": "https://x.com/conscious_seeker/status/1234567890123456789",
  "media": [],
  "user_id": "987654321",
  "user_verified": false,
  "tweet_id": "1234567890123456789",
  "is_reply": false,
  "is_quote": false,
  "conversation_id": "1234567890123456789",
  "in_reply_to_user_id": null,
  "language": "en",
  "in_reply_to_username": null,
  "quoted_tweet_id": null,
  "user_display_name": "Conscious Seeker",
  "datetime": "2025-10-24T19:03:40+00:00"
}

Data Splits

The dataset is provided as a single unsplit collection for maximum flexibility. We recommend creating custom splits, e.g.:

Train/Validation/Test: 80/10/10 random split. Temporal Split: Use datetime to split by time periods (e.g., pre-2025 vs. post-2025).

To split in code:

from datasets import load_dataset
import random

dataset = load_dataset("InnerI/consciousness-hashtag")
# Random split example
dataset = dataset["train"].train_test_split(test_size=0.2, seed=42)

Dataset Creation

Curation Rationale Curated to mirror the evolving public fascination with consciousness—fueled by AI advancements, quantum theories, and spiritual revivals—this dataset distills X's raw energy into a structured resource. By targeting "consciousness" or #consciousness, it captures diverse voices, enabling studies on how global events (e.g., 2025 AI ethics summits) shape collective awareness.

Source Data

• Platform: X (formerly Twitter).
• Collection Method: Scraped via X API using keyword/hashtag filters ("consciousness" OR "#consciousness"), deduplicated, and limited to English-language tweets. No manual curation; automated filtering for relevance.
• Time Period: Spans up to October 2025, with timestamps in datetime for precise tracking.
• Size: 988 tweets, post-preprocessing (e.g., removal of spam/low-engagement posts).
• Preprocessing: Basic cleaning (e.g., URL normalization, hashtag extraction); no advanced annotation.

Annotations

Raw data with no pre-applied labels. Fields like label are reserved for user annotations (e.g., via tools like Prodigy or LabelStudio).

Considerations for Using the Data

Social Impact of Dataset: This dataset democratizes access to consciousness discourse, supporting research in AI alignment, mental health, and cultural anthropology. It can inform ethical AI development by highlighting public concerns (e.g., machine consciousness fears) and foster inclusive dialogues on awareness.

Discussion of Biases

• Platform Bias: Overrepresents X's user base (e.g., tech-savvy, English-speaking demographics), potentially underrepresenting non-Western or marginalized voices.
• Content Bias: Skews toward trending or vocal opinions (e.g., panpsychism surges during viral threads); scientific tweets may outnumber empirical ones.
• Temporal Bias: Recent data (2024–2025) dominates due to API recency, influenced by events like AI breakthroughs.
• Mitigation: Use stratified sampling by datetime or user_verified; diversify with external datasets for balance.

Other Known Limitations

• Null Fields: ~20% of metadata (e.g., user_id, language) may be null due to API restrictions or privacy deletions—handle with imputation or filtering.
• Noise: Includes tangential tweets (e.g., "consciousness in marketing"); apply relevance scoring for precision.
• Size Constraints: Substantial but not exhaustive; for larger scales, combine with similar datasets.
• Privacy/Ethics: Usernames/IDs are anonymized where possible; avoid re-identification in publications.

Additional Information

Dataset Curators

innerinetcompany.com & macrocosmos.ai: Pioneers in AI-curated explorations of consciousness and cosmic data. Focused on bridging human intuition with machine insight.

Licensing

Released under the CC BY 4.0 license (Creative Commons Attribution 4.0 International). Free for research, commercial, and educational use, with attribution to InnerI/macrocosmos.ai. Citation Information Please cite this dataset as:

@dataset{InnerI_2025,
  author = {InnerI and macrocosmos.ai},
  title = {consciousness-hashtag: A 100K Tweet Dataset on Consciousness Discourse},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/InnerI/consciousness-hashtag}
}

Contributions

Open to enhancements! Submit issues/pull requests on the Hugging Face repo or tag @innerinetco on X. Ideas: Add pre-computed embeddings, multilingual expansions, or bias audits.

Getting Started

Load and explore with the Hugging Face datasets library:

from datasets import load_dataset
import pandas as pd

# Load the dataset
dataset = load_dataset("InnerI/consciousness-hashtag")

# Quick peek
print(dataset["train"][0])  # First tweet

# Analyze basics with pandas
df = pd.DataFrame(dataset["train"])
print(df["text"].head())  # Sample texts
print(df["datetime"].describe())  # Temporal overview

# Example: Filter for recent tweets
recent = df[df["datetime"] > "2025-01-01"]
print(recent["text"].str.contains("#consciousness").sum())  # Count with hashtag

Dive deeper: What thread of consciousness calls to you? Share your analyses with #consciousness-hashtag 🚀