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.
-
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
- 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
We use BU's Shared Computing cluster to run the graoh generation scripts.
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.shYou 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
- Now, using
asl-playground.ipynbfile 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/
- 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.
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.