README.md

---
license: cc-by-4.0
task_categories:
- text-classification
- time-series-forecasting
language:
- en
- multilingual
tags:
- google-trends
- trending-now
- attention-dynamics
- information-diffusion
- temporal-analysis
- search-trends
size_categories:
- 1M<n<10M
pretty_name: Google Trend Archive
---

# Google Trend Archive: Global Real-Time Search Trends (2024-2026)

## Dataset Details

### Dataset Description

This dataset contains over **7.6 million trending search instances** from Google's Trending Now feature, collected continuously from November 28, 2024 to January 3, 2026 across all available geographic locations (200+ countries/regions). Unlike aggregated retrospective tools like Google Trends, Trending Now captures search queries experiencing real-time surges, offering unprecedented temporal granularity for studying collective attention dynamics.

Each instance represents a moment when specific search terms or query clusters became "trending" according to Google's algorithm, complete with search volume buckets, precise timestamps, trend durations, geographic locations, and related query variations.

- **Curated by:** Aleksandra Urman, Anikó Hannák, Joachim Baumann (Social Computing Group, University of Zurich & Stanford Artificial Intelligence Laboratory)
- **Language(s):** Multilingual
- **License:** CC-BY-4.0
- **DOI:** https://doi.org/10.57967/hf/7531

### Dataset Sources

- **Repository:** https://huggingface.co/datasets/aurman/GoogleTrendArchive
- **Contact:** urman@ifi.uzh.ch

## Uses

### Direct Use

This dataset enables research across multiple domains:

- **Information Diffusion Modeling**: Track how topics cascade across geographic regions and analyze diffusion pathways
- **Event Detection**: Identify breaking news, crises, and significant events from search surge patterns  
- **Comparative Cultural Studies**: Analyze differences in collective attention across countries and cultures
- **Crisis Communication**: Understand information needs during emergencies
- **Temporal Pattern Analysis**: Study daily, weekly, and seasonal rhythms in collective attention
- **Predictive Modeling**: Develop models to forecast trend emergence, duration, and spread
- **Media Ecosystem Analysis**: Compare search trends with social media and news coverage
- **Political Communication**: Examine attention to political topics across different political systems

### Out-of-Scope Use

The dataset should **not** be used for:

- **Surveillance or monitoring** of specific populations
- **Marketing or commercial targeting** (intended for academic research only)
- **Drawing deterministic conclusions** about population beliefs (search shows information-seeking, not settled beliefs)
- **Identifying or tracking individuals** (data is aggregated, but de-anonymization attempts would violate ethical norms)
- **Training models for harmful applications** (discrimination, manipulation, surveillance)

## Dataset Structure

Each instance contains six fields:

1. **Trend Identifier**: The search query or representative term for a cluster of related queries (e.g., "man united vs bodø/glimt")
2. **Search Volume**: Bucketed traffic ranges (e.g., "100+", "50K+", "2M+") indicating approximate search volume
3. **Start Timestamp**: When Google's system first flagged the trend as emerging (ISO format with timezone)
4. **End Timestamp**: When the trend ended and returned to baseline (ISO format with timezone)
5. **Trend Breakdown**: Comma-separated list of related query variations that Google clustered together
6. **Explore Link**: URL to the corresponding Google Trends page for further investigation

**Total Instances**: 7,600,000+ trending searches  
**Temporal Coverage**: November 28, 2024 - January 3, 2026 (with ~14 days of gaps due to technical issues)  
**Geographic Coverage**: 1358 countries and regions  
**Format**: CSV files with UTF-8 encoding

## Dataset Creation

### Curation Rationale

This dataset is a comprehensive archive of Google Trending Now data spanning over one year (from November 28, 2024 to January 3, 2026) across all available geographic locations. Unlike aggregated retrospective tools like Google Trends, Trending Now captures search queries experiencing real-time surges, offering temporal granularity for studying collective attention dynamics --- however, this data is not available retrospectively through Google directly. Our dataset addresses this gap by presenting an archive of Trending Now data.

### Source Data

#### Data Collection and Processing

Data was collected using automated software that continuously monitored Google Trends Trending Now pages (e.g., https://trends.google.com/trending?geo=US) for all available geographic locations from November 28, 2024 onward.

**Validation procedures included:**
- Automated checks for data completeness across time periods and locations
- Verification of timestamp consistency  
- Deduplication to handle multiply-collected trends
- Comparison against manually spot-checked trends
- Logging of collection failures (~14 days of gaps)

**Preprocessing steps:**
- Timestamp standardization to ISO 8601 format with timezone information
- Duration calculation from start/end timestamps
- Geographic code standardization
- Format consolidation from daily files into unified dataset
- Data validation and quality checks

All preprocessing code and raw data files are available in the repository for transparency and replication.

#### Who are the source data producers?

Google's Trending Now system, which aggregates and anonymizes search activity from Google users worldwide. 

### Personal and Sensitive Information

**No personally identifiable information is included.** 

## Bias, Risks, and Limitations

**Technical Limitations:**
- Search volumes provided in buckets rather than exact counts
- Google's trend identification algorithm is proprietary (clustering and thresholding decisions not transparent)
- ~14 days of missing data due to technical collection issues
- Trends represent relative surges, not absolute search volumes
- Some trends may have estimated rather than actual end dates

**Representativeness Issues:**
- Search behavior does not uniformly represent entire populations
- Digital divides and varying internet penetration mean certain demographics are over/underrepresented  
- Search engine market share varies by country (Google dominance differs across regions)
- Geographic coverage limited to locations where Google provides Trending Now data

**Algorithmic Mediation:**
- Google determines what qualifies as "trending" using proprietary algorithms
- Content filtering based on Google's policies may exclude certain trends
- Reflects Google's selection of notable search spikes, not all search activity

**Interpretation Risks:**
- Search trends show what people search for, not why they search or what they conclude
- Risk of drawing inappropriate generalizations or stereotypes about populations
- Patterns in sensitive topics (health, politics) could stigmatize communities
- Missing social context can lead to misinterpretation

### Recommendations

Users should:
- Acknowledge representativeness limitations in cross-country comparisons
- Account for structural inequalities in internet access when interpreting patterns
- Avoid causal claims without additional evidence
- Exercise caution when analyzing/presenting sensitive topics
- Consider social context when interpreting collective attention patterns
- Present findings with appropriate epistemic humility
- Recognize this represents one platform's perspective on collective attention

## Citation

**BibTeX:**
```bibtex
@dataset{urman2026googletrendarchive,
  title={Google Trend Archive: Global Real-Time Search Trends},
  author={Urman, Aleksandra and Hann{\'a}k, Anik{\'o} and Baumann, Joachim},
  year={2026},
  publisher={Hugging Face},
  doi={10.57967/hf/7531},
  url={https://huggingface.co/datasets/aurman/GoogleTrendArchive}
}
```

**APA:**

Urman, A., Hannák, A., & Baumann, J. (2026). *Google Trend Archive: Global Real-Time Search Trends* [Dataset]. Hugging Face. https://doi.org/10.57967/hf/7531

## More Information

### Funding

Research activities during dataset construction received support from:
- Swiss National Science Foundation – PostDoc Mobility fellowship P500-2 235328 (JB)
- SNSF Project Grant 215354 (AU and AH)

### Maintenance

The dataset is hosted on Hugging Face and maintained by the authors. The current release represents a snapshot from November 28, 2024 to January 3, 2026. **Updates are planned every 2-3 months** as data collection is ongoing, communicated through changelogs in the repository.

All versions will be permanently archived and accessible through Hugging Face's versioning system to ensure reproducibility.

### Contributions

We welcome community contributions:
- **Derived datasets**: Researchers adding annotations or features are encouraged to share as separate datasets with attribution
- **Corrections**: Report errors to urman@ifi.uzh.ch for review and incorporation into updated versions
- **Extensions**: Researchers collecting additional data using our methodology may collaborate on future releases

## Dataset Card Authors

Aleksandra Urman, Anikó Hannák, Joachim Baumann

## Dataset Card Contact

urman@ifi.uzh.ch