---
license: mit
task_categories:
- text-classification
- time-series-forecasting
language:
- en
tags:
- github
- trending
- repositories
- software-engineering
- popularity
- time-series
size_categories:
- 100K<n<1M
---

# GitHub Trending Projects (2013-2025)

A comprehensive dataset of **423,098 GitHub trending repository entries** spanning **12+ years** (August 2013 - November 2025), scraped from Wayback Machine snapshots of GitHub's trending page.

## 🎯 Dataset Overview

This dataset captures the evolution of GitHub's trending repositories over time, providing insights into:
- **Software development trends** across programming languages and domains
- **Popular open-source projects** and their trending patterns
- **Community interests** and shifts in developer focus over 12 years
- **Viral repository dynamics** and sustained popularity patterns

**Key Statistics:**
- 📊 **423,098** trending repository entries
- 🗂️ **14,500** unique repositories
- 📅 **128 months** of coverage (2013-08 to 2025-11)
- ⭐ **89.8%** scraping success rate from Wayback Machine
- 🏆 **Pre-processed monthly rankings** with weighted scoring

## 🔧 Dataset Configurations

This dataset has **two configurations**:

### Configuration: `full` (Default)
Complete daily trending data with 423,098 entries

```python
from datasets import load_dataset
ds = load_dataset('ronantakizawa/github-top-projects', 'full')
```

### Configuration: `monthly`
Top 25 repositories per month with 3,200 entries

```python
from datasets import load_dataset
ds = load_dataset('ronantakizawa/github-top-projects', 'monthly')
```

## 📁 Dataset Files

### 1. `github-trending-projects-full.csv` (19 MB)
**Complete daily trending data** - All 423,098 entries

| Column | Type | Description |
|--------|------|-------------|
| `name` | string | Repository name |
| `star_count` | integer | Star count (max recorded, may be empty for pre-2020) |
| `fork_count` | integer | Fork count (max recorded, may be empty for pre-2020) |
| `repo_owner` | string | Repository owner/organization |
| `rank` | integer | Position in trending (1-25) |
| `date` | date | Snapshot date (YYYY-MM-DD) |

**Use cases:**
- Custom time-series analysis
- Training ML models on trending patterns
- Analyzing daily trending dynamics
- Creating custom aggregations (weekly, yearly, etc.)
- Studying viral repository behavior

### 2. `github-top-projects-by-month.csv` (211 KB)
**Monthly top 25 repositories** - Pre-processed with weighted scoring

| Column | Type | Description |
|--------|------|-------------|
| `month` | string | Month (YYYY-MM) |
| `rank` | integer | Monthly rank (1-25) |
| `repository` | string | Full repository name (owner/name) |
| `repo_owner` | string | Repository owner |
| `repo_name` | string | Repository name |
| `star_count` | integer | Maximum recorded stars |
| `fork_count` | integer | Maximum recorded forks |
| `ranking_appearances` | integer | Times appeared in trending that month |

**Use cases:**
- Quick monthly insights and visualizations
- Dashboard creation
- Identifying consistently popular projects
- Trend analysis without processing overhead

## 🏆 Scoring Methodology

Monthly rankings use a **weighted frequency and position-based scoring system**:

```
Score = Σ (25 - rank + 1) for each trending appearance

Where:
- Rank 1 → 25 points
- Rank 2 → 24 points
- ...
- Rank 25 → 1 point
```

**Example:**
- Project appears 10 times at rank #1 → 250 points
- Project appears 20 times at rank #10 → 320 points (higher ranked!)

This rewards both **consistency** (frequent appearances) and **position** (higher ranks).

## 📊 Data Collection

**Source:** GitHub Trending page via Wayback Machine (web.archive.org)
**Period:** August 21, 2013 - November 30, 2025
**Method:** Python web scraping with BeautifulSoup
**Snapshots:** 17,127 successfully scraped from 19,064 available
**Retry Logic:** Up to 15 retries with exponential backoff

**HTML Parsing:**
- Multiple extraction methods for different page structures (2013-2019, 2020+)
- Handles changes in GitHub's trending page design
- Robust error handling for incomplete snapshots

## ⚠️ Known Limitations

### 1. **Missing Star/Fork Data (Pre-2020)**
- **100% of 2013-2019 entries** lack star/fork counts
- Only **67.8% of dataset** has popularity metrics
- Cause: Different HTML structure in older Wayback snapshots
- **Impact:** Cannot compare absolute popularity for historical projects

### 2. **Uneven Temporal Distribution**
- Snapshot frequency: **1 to 31 per month** (31x variance)
- 2019-2020 heavily over-represented
- Some months: 25 projects, others: 17,446 projects
- **Impact:** Monthly scores favor periods with more snapshots

### 3. **Star Count Timing Inconsistency**
- Star counts are "maximum ever recorded" across all snapshots
- A 2015 project's stars might be from 2025 scraping
- **Not temporally aligned** - older projects had more time to accumulate stars
- **Impact:** Can't fairly compare popularity across eras

### 4. **Multiple Appearances Bias**
- Top projects appear 1,700-1,900 times
- 1,129 projects appear only once
- Favors "evergreen" educational repos
- **Impact:** Brief viral projects may be undervalued

### 5. **Failed Scrapes**
- 1,937 URLs failed (10.2%)
- Mainly 2014-2019 due to SSL/TLS incompatibility
- Some date ranges completely missing
- **Impact:** Gaps in temporal coverage

**See DATASET_ISSUES.md for detailed analysis**

## 📈 Data Quality by Era

| Era | Quality | Star Data | Snapshot Density | Grade |
|-----|---------|-----------|------------------|-------|
| **2013-2019** | Limited | ❌ 0% | Low-Medium | **C+** |
| **2020-2025** | Excellent | ✅ 100% | High | **A-** |

**Recommendation:** Use 2020-2025 data for analyses requiring star/fork counts

## 💡 Usage Examples

### Load with Hugging Face Datasets (Recommended)

```python
from datasets import load_dataset

# Load complete daily dataset (423,098 entries)
ds_full = load_dataset('ronantakizawa/github-top-projects', 'full')
df_full = ds_full['train'].to_pandas()

# Load monthly top 25 dataset (3,200 entries)
ds_monthly = load_dataset('ronantakizawa/github-top-projects', 'monthly')
df_monthly = ds_monthly['train'].to_pandas()

# Filter to 2020+ (with star data)
df_recent = df_full[df_full['date'] >= '2020-01-01']

# Get November 2025 top 10
nov_2025 = df_monthly[df_monthly['month'] == '2025-11'].head(10)
print(nov_2025[['rank', 'repository', 'star_count', 'ranking_appearances']])
```

### Load Directly from CSV

```python
import pandas as pd

# Download files from the dataset page, then:
df_full = pd.read_csv('github-trending-projects-full.csv')
df_monthly = pd.read_csv('github-top-projects-by-month.csv')

# Get top trending projects of 2024
df_2024 = df_full[df_full['date'].str.startswith('2024')]
top_2024 = df_2024.groupby(['repo_owner', 'name']).size().sort_values(ascending=False).head(10)
print(top_2024)
```

### Time Series Analysis

```python
import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv('github-trending-projects-full.csv')
df['date'] = pd.to_datetime(df['date'])

# Analyze a specific project over time
project = 'microsoft/vscode'
project_df = df[(df['repo_owner'] == 'microsoft') & (df['name'] == 'vscode')]

# Plot trending frequency over time
monthly_counts = project_df.groupby(project_df['date'].dt.to_period('M')).size()
monthly_counts.plot(title=f'{project} Trending Frequency')
plt.ylabel('Days in Trending')
plt.show()
```

### Language Trends (requires additional metadata)

```python
# Note: Language data not included in this dataset
# Would need to join with GitHub API data or another dataset
```

## 🔍 Research Applications

This dataset enables analysis of:

1. **Trending Dynamics**
   - What makes a repository go viral?
   - How long do projects stay trending?
   - Seasonal patterns in software development

2. **Technology Adoption**
   - Rise and fall of programming languages
   - Framework popularity over time
   - Shift from monolithic to microservices

3. **Open Source Evolution**
   - Growth of educational repositories
   - Corporate open source contributions
   - Regional trending patterns

4. **Predictive Modeling**
   - Forecasting future trending projects
   - Identifying early viral indicators
   - Star growth prediction models

5. **Developer Behavior**
   - Community interest shifts
   - Popular project categories
   - Documentation and tutorial demand

## 📊 Example Insights

**Most Consistently Trending (2020-2025):**
1. `jwasham/coding-interview-university` - 1,948 appearances
2. `TheAlgorithms/Python` - 1,891 appearances
3. `donnemartin/system-design-primer` - 1,865 appearances

**Trending Patterns:**
- Educational repositories dominate long-term trending
- AI/ML projects saw massive spike in 2023-2024
- Web frameworks remain consistently popular

## 📜 License

**MIT License**

This dataset is released under the MIT License. You can:
- ✅ Use for commercial purposes
- ✅ Modify and distribute
- ✅ Use in research (attribution appreciated!)
- ✅ Include in proprietary software

**Source Data:** Wayback Machine (public archive)
**Original Content:** GitHub Trending pages

## 🙏 Acknowledgments

- **GitHub** for maintaining the trending page
- **Internet Archive** for the Wayback Machine
- **Open Source Community** for creating amazing projects

## 📧 Contact & Contributions

- **Issues/Questions:** Open an issue on the dataset repository
- **Data Errors:** Please report any inconsistencies
- **Contributions:** Additional metadata or corrections welcome

## 📖 Citation

If you use this dataset in your research, please cite:

```bibtex
@dataset{github_trending_2013_2025,
  title={GitHub Trending Projects Dataset (2013-2025)},
  author={Your Name},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/YOUR_USERNAME/github-top-projects}
}
```

## 🔗 Related Datasets

- GitHub Archive (gharchive.org) - Complete GitHub event stream
- GHTorrent - GitHub data for research
- Libraries.io - Package manager dependency data

## 📅 Updates

- **2025-12:** Initial release (2013-08 to 2025-11)
- Future updates planned quarterly

## ⚙️ Technical Details

**Scraping Configuration:**
- Retry attempts: 15 with exponential backoff
- Delay between requests: 4-6 seconds (randomized)
- Timeout: 45 seconds per request
- User-Agent: Mozilla/5.0 (standard browser)

**Data Processing:**
- Deduplication: By date + repo + rank
- Sorting: Chronological (newest first)
- Encoding: UTF-8
- Format: CSV with headers

## 🐛 Known Issues

See `DATASET_ISSUES.md` for comprehensive list including:
- Missing data gaps
- Star count timing issues
- Temporal distribution variance
- Recommended usage guidelines

---

**Last Updated:** December 2025
**Dataset Version:** 1.0
**Status:** ✅ Complete and ready for use