Skip to content

poloclub/argo-scholar

Repository files navigation

Argo Scholar

Build Status arxiv badge GitHub

An interactive literature exploration visualization system that runs in your web browsers. No installation needed.

Argo Scholar logo

Launch Argo Scholar in your browser

Quick Start

For beginners, Argo Scholar provides 2 built-in sample paper citation and reference networks for literature exploration: Deep Learning and Apolo Sample.

If you have launched Argo Scholar with the above link, Deep Learning will automatically be displayed. You can also load sample networks at any time by selecting Graph -> Load Sample in the top menu.

Basic navigation

Whether you are on mouse/keyboard or a touchscreen device, you can learn the basic navigation using the Help (?) button on the top right corner of the app. This will help you learn how to pan, zoom and select nodes.

If you are on a device with small screen and cannot see the top menu, you can click the expand button to bring up the full Argo Scholar UI.

Force-directed Layout

Once you have launched the sample network, you will see a play/pause button on the top menu bar. This is for interactive force-directed layout, which helps to position your nodes.

Argo Scholar visualization with force directed layout

Graph Options Panel

Argo Scholar gives a default visualization by coloring and sizing the nodes based on their PageRank or Degree values. You can update these settings using the panel on the left (when you are not selecting any node).

Argo Scholar visualization graph options

Override Individual Nodes

If you select a node, you will see the Graph Options Panel changed into override mode. You can override the global settings by giving these selected nodes a different look!

Argo Scholar override options

Pinning and Unpinning

If you want to fix the positions of certain nodes when other nodes are running force-directed layout, you can select a node or a group of nodes and use the Pin button on the selection menu that pops up. (By default, if you select a node and drag it to a new position, it will already be pinned).

To unpin, just select them again and click the Unpin button.

Exploring Your Own Network

In order to build your network from the scratch, you may want to start with a blank canvas. Argo Scholar allows you to start your own literature exploration process from an empty graph at any time by simply selecting Graph -> New...

Adding Nodes via Search

To add new paper nodes, you can utilize the search bar on the top. Argo Scholar allows search by keyword or Corpus ID. You can do this to add as many papers as you so wish yo your network graph.

NOTE: You can look up the Corpus ID of a paper through Semantic Scholar. The CorpusID will be located on the paper's Semantic Scholar page, next to the paper title.

Incrementally Adding Neighbors

Now we have starting paper nodes, we can add some neighbor nodes!

Right-click the paper you find interesting. You will see the option to either Add 5 Paper Citations or Add 5 Paper References. Clicking either option will add 5 addition papers of either category to the literature network. You can keep on adding neighbors until you have added all possible papers.

NOTE: As defined on Semantic Scholar, citations refer to the papers that have cited the current paper, and references refer to the papers were referenced in the current paper.

After adding the citations or references, you can now explore the newly added nodes. If you are not find certain papers interesting or relevant, you can hide them from view by clicking on the paper and selecting Hide.

Now select 1 paper that you find interesting, you will see a Neighbors (xx nodes hidden) button on the floating selection menu. Clicking on it will bring you to the Neighbor Menu.

In the Neighbor Menu, you can see a table listing all the neighbor nodes and their attributes. You can sort them by their attributes, individually add or hide them in the network, or use the tools above to add an arbitrary number of neighbors with top PageRank or Degree values. This helps you identify highly cited papers that has cited or is cited by the current selected paper that you are interested in.

Argo Scholar incremental exploration

If you are using a mouse (as opposed to a touchscreen device), you can find a shortcut to do the same thing by right clicking a node.

You can also individually manage nodes in Tools -> Data Sheet.

When you are done, save or share your snapshot using the Graph menu!

Saving and Sharing

Network Snapshots

Argo saves your visualization and exploration progress into snapshots. A snapshot includes the full network data (including nodes and connections) as well as the current visualization settings.

You can capture a snapshot using the Graph -> Save Snapshot for saving locally, or Graph -> Publish and Share Snapshot for saving your snapshot to a URL/link. You can also rename the snapshot by clicking on its name (Untitled Graph by default) on the menu bar.

If you have saved your snapshot, you can import the network at anytime in the future and resume right where you left off.

You can find sample files to import in the samples directory of this repository.

Now try using Graph -> Open Snapshot.

Sharing as links/URLs

Now try Graph -> Publish and Share Snapshot.

By sharing your literature network with a link, anyone can load the network through the link later. It's a great tool for sharing and collaboration.

Argo Scholar sharing graph as link

Sharing as Embedded Widgets

On the same screen where you get your sharable URL, you can also copy the iframe code for embedding the snapshot. Argo Scholar allows you to embed any snapshot URL in iframes. This is perfect for publishing your literature network on online articles, blog posts or interactive notebooks (such as a Jupyter Notebook).

Argo Scholar embedded widget mode

About Sharing Service

We provide a public sharing service for public datasets. If you want to establish your own sharing server for private or proprietary datasets, refer to the Deployment Guide to easily set up your own sharing service!

Documentations

If you want to learn about the development process or how to deploy your own version of Argo Scholar, please check out the following documentations:

Credits

♥ Argo Scholar was developed and maintained by Kevin Li, Alex Yang, Anish Upadhayay, Zhiyan Zhou, Jon Saad-Falcon, Duen Horng Chau from Polo Club of Data Science at Georgia Tech.

Citation

@inproceedings{li2022argoscholar, 
  author = {Li, Kevin and Yang, Haoyang and Montoya, Evan and Upadhayay, Anish and Zhou, Zhiyan and Saad-Falcon, Jon and Chau, Duen Horng},
  title = {Visual Exploration of Literature with Argo Scholar},
  year = {2022},
  isbn = {9781450392365},
  publisher = {Association for Computing Machinery},
  url = {https://doi.org/10.1145/3511808.3557177},
  doi = {10.1145/3511808.3557177}
}

License

Argo Scholar is available under the MIT License. Argo Scholar uses the Semantic Scholar Open Research Corpus API, which is licensed under ODC-BY. More can be found here: Waleed Ammar et al. 2018. Construction of the Literature Graph in Semantic Scholar. NAACL

Contact

If you have any questions or would like to learn more about the project, feel free to contact Kevin Li or Alex Yang.