ScholTrack: Automate Finding and Fetching Academic Citations

Online Tool (Search by Paper IDs) | Demo

ScholTrack is a Python tool designed to retrieve and save citations to a specified paper (or a list of papers) into a single file. It is especially useful for surveying a research field by gathering papers that cite key works.

The tool uses the public Semantic Scholar API to collect citations and extract important details such as titles, abstracts, citation counts, publishing venues, arXiv links, and more. You can save the data in formats like CSV, JSON, or TXT for further for further analysis or integration with research and LLM tools. ScholTrack can be used via the command line interface or through a simple Colab-based online tool.

Use Cases

📚 Fetch all citations to a specific paper
Retrieve all citations to a specified paper and save them to a CSV file.
📘 Demo: Fetch citations for 3D Gaussian Splatting (3DGS)
🗞️ Get the latest works in a research field
Find the most recent papers by collecting citations to a set of seminal works, sorted by publication date.
📘 Demo: Fetch the latest works in Novel View Synthesis
🧩 Find papers that combine methods X and Y
Discover papers that cite seminal works from different fields, showing papers that combine methods X and Y.
📘 Demo: Fetch papers that combine 3DGS and SMPL for human avatar modeling

Installation

pip install scholtrack

Getting Started

To use ScholTrack, simply find the Semantic Scholar Paper IDs for the papers you are interested in and run the following command:

scholtrack -p 2cc1d857e86d5152ba7fe6a8355c2a0150cc280a -o 3dgs_citations.csv

This command retrieves all papers citing 3D Gaussian Splatting and saves the results in a CSV file (3dgs_citations.csv). The file includes abstracts, citation counts, arXiv links, publication venues, and more.

Online Tool

For a simple, no-code demo, try the Colab online tool. This tool lets you input a list of Semantic Scholar Paper IDs and retrieve the corresponding citation data.

Examples

Example 0: Fetch All Citations for a Paper and Save to CSV

This basic command retrieves all citations and related information (such as abstract, arXiv link, citation count, venue, etc.) for a given paper and saves it to a CSV file:

scholtrack -p 2cc1d857e86d5152ba7fe6a8355c2a0150cc280a -o 3dgs_citations.csv

Example 1: Fetch Latest Citations to a Paper, Save to TXT

You can order the output based on the paper arXiv ID to prioritize latest relevant publications:

scholtrack -p 2cc1d857e86d5152ba7fe6a8355c2a0150cc280a -o 3dgs_citations.txt -s arxiv -t txt

Example 2: Fetch All Citations to a List of Papers

You can provide more than one Paper ID either directly in the command line:

scholtrack -p 2cc1d857e86d5152ba7fe6a8355c2a0150cc280a 428b663772dba998f5dc6a24488fff1858a0899f -o nvs_citations.txt

or by providing a TXT file with the Paper ID list:

scholtrack -f nvs.txt -o nvs_citations.txt

Here, nvs.txt is the following file which contains recent seminal works in the novel view synthesis (NVS) field:

# Collection of Semantic Scholar Paper IDs related to Neural Rendering, NeRF, and 3DGS
428b663772dba998f5dc6a24488fff1858a0899f # Nerf: Representing scenes as neural radiance fields for view synthesis 
21336e57dc2ab9ae2171a0f6c35f7d1aba584796 # Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields
2cc1d857e86d5152ba7fe6a8355c2a0150cc280a # 3D Gaussian Splatting for Real-Time Radiance Field Rendering

You can create custom paper collections to track daily progress in your field of interest or conduct comprehensive surveys.

Note: Retrieving results for papers with thousands of citations or long citation lists may take a few minutes. Additionally, the current limit for citations that can be retrieved from the Semantic Scholar API without an API key is 10,000 per paper.

Example 3: Finding Papers at the Intersection of Fields

You can perform a more targeted search to find papers that cite multiple works from a given list, which is useful for identifying research at the intersection of different fields. For example, the command below retrieves all papers that cite both 3D Gaussian Splatting and the seminal work "SMPL: A Skinned Multi-Person Linear Model" on digital humans. This provides a comprehensive view of how 3DGS techniques are applied in digital avatar modeling:

scholtrack -o avatar_3dgs.csv -n 2 -p 2cc1d857e86d5152ba7fe6a8355c2a0150cc280a 32d3048a4fe4becc7c4638afd05f2354b631cfca

Here, -n 2 specifies that only papers citing at least two works from the list will be included.

Example 4: Using Pre-built Paper Collections

You can also explore pre-packaged lists of papers included in the repository. Currently, we offer the "Novel View Synthesis" (nvs) collection, which tracks papers related to this topic, including works such as NeRF, 3DGS, and DreamFusion. To retrieve the corresponding citations, run:

scholtrack -c nvs -o novel_view_synthesis_works.csv -s arxiv

In this example, -s arxiv sorts the results by the arXiv date of appearance.

Finding Paper IDs

ScholTrack uses Semantic Scholar Paper IDs to retrieve citations. To find a Paper ID:

Go to Semantic Scholar.
Search for the paper.
Click on the paper in the results.
The Paper ID is the string after the last slash in the URL, e.g., for 3D Gaussian Splatting:
```
https://www.semanticscholar.org/paper/3D-Gaussian-Splatting-for-Real-Time-Radiance-Field-Kerbl-Kopanas/2cc1d857e86d5152ba7fe6a8355c2a0150cc280a
```
The Paper ID is 2cc1d857e86d5152ba7fe6a8355c2a0150cc280a.

Use this ID in ScholTrack to retrieve citations. Alternatively, you can provide the Semantic Scholar URL directly as input:

scholtrack -u https://www.semanticscholar.org/paper/3D-Gaussian-Splatting-for-Real-Time-Radiance-Field-Kerbl-Kopanas/2cc1d857e86d5152ba7fe6a8355c2a0150cc280a -o 3dgs_references.csv

Python API Usage

from scholtrack.api import CitationExplorerAPI
from scholtrack.exporter import CitationExporter

# Initialize the API client (API key is optional)
api = CitationExplorerAPI()

# Fetch citations for a paper
paper_ids = ["395bfdae59f1a5fde16213cade43d2587e4565df",
             "8ba2c82fe675ede1d5c12fc4f97cf8ca3ebf1ca3"]

citations = api.get_citations_for_papers(paper_ids)

# Save citations to a CSV file
CitationExporter.save_to_csv(citations, filename="citations.csv")

This will save the retrieved citation data for the specified Paper IDs to a CSV file.

Development version

For the latest development version, install directly from GitHub:

pip install git+https://github.com/sergeyprokudin/scholtrack.git

License

This project is licensed under the Apache License 2.0. See the LICENSE file for more information.

Contributions

Contributions are welcome. Feel free to submit pull requests, report issues, or suggest new features.

Acknowledgments

This tool leverages the Semantic Scholar API to fetch scholarly citation data. Special thanks to the Semantic Scholar team for providing this valuable service.

sergeyprokudin / scholtrack

readme