ASL-LEX

About

ASL-LEX is a database of lexical and phonological properties of nearly 3,000 signs in American Sign Language. This project is a collaboration between the Laboratory for Language and Cognitive Neuroscience at San Diego State University and the Psycholinguistics and Linguistics Lab at Tufts University. ASL-LEX is a searchable database of subjective frequency ratings, iconicity ratings, lexical properties (e.g., initialized signs; lexical class), and six phonological features from which neighborhood densities have been calculated. ASL-LEX also provides, reference video clips, English translations and, for a subset of signs, alternative translations.

Accessing the ASL-LEX 2.0 web interface

• The visualization is currently hosted on Github Pages.

• View the older interface at the ASL 1.0 Link

• It consists of two visualizations, a Network Graph that depicts the lexicon as a network graph where nodes are connected if they have phonological and lexical similarities. Hover on a node to view it's neighbors

• Click on it for more information about that sign.

• You can also filter the network by applying various filtering criteria.

• Hover over any property to view more info about that property

• Hover over a sign to view more quick info about that sign

• There is also a Pair Plots visualization which depcits distribution of signs across various pairs of properties such as Lexical Class, Neighborhood Density, Iconicity and Frequency.

• You can drag to select a subset of signs on either visualization and view the subset in the other view.

• For more ways to explore the data, please reach out to the team at hicsail@bu.edu

Custom Development

Starting the project locally

• Clone the repo and in the root folder, start the server using MAMP or python Simpleserver
• Go to http://localhost:8000/visualization to view the interface

Running the script to generate the network graph and preparing a new visualization

We use BU's Shared Computing cluster to run the graoh generation scripts.

Step 1

To run it on the SCC:

• login into a compute node
• cd into /project/hariri/asl-lex/
• Run make-pipeline.sh using qsub
• Note that by default, it runs graph creation with the default edge criterion: qsub -o /project/hariri/asl-lex/logs/ -N RUN_PYND /project/hariri/asl-lex/data-analysis/scripts/default-job.sh You can change this by using another edge criterion file from the scripts folder.
• The pipeline will provide files to be used for the next step of the visualization under the generated-data folder, once it is complete. Please refer to the BU SCC documentation on how to check logs and access the status of your running job: http://www.bu.edu/tech/support/research/system-usage/scc-quickstart/#BATCH

Step 2

• Now, using asl-playground.ipynb file under notebooks, run the notebook to obtain the JSON files for the visualization, namely graph.json, constraints.json and sign-props.json
• Copy these over to the visualization/data folder, removing older files with the same name
• Run the D3 Force Layout Visualization by going to http://localhost:8000/visualization/force-layout-rendering.html.
• Wait for the graph to stablize and download the graph.json file using "Download Coordinates File" link, now with x,y coordinates attached with each node. Wait at least 5 minutes for the graph to stabilize.
• Replace the graph json file with the new one obtained with coordinates in the previous step.
• View the updated visualization at http://localhost:8000/visualization/

Deployment

• Once changes are done, remove your local copy of gh-pages branch. Then checkout a new gh-pages branch from the branch you made changes on.
• Remove everything except the visualization folder.
• Move all files from visualization folder to the root level.
• Push to the remote gh-pages branch from your local gh-pages branch.
• The new interface will be up on the Github Pages Link in a few minutes.

License

The content on this website is available under a CC-By-NC license, meaning you can reuse and remix this content with attribution for non-commercial purposes. If you would like to cite it, please use the following: Caselli, N., Sevcikova, Z., Cohen-Goldberg, A., Emmorey, K. (in prep). ASL-LEX: A Lexical Database for ASL.

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.