Researchers in the field of linguistics are interested in discovering interesting non-random feature correlations across languages that are not historically or geographically related. A feature is a property of a language that has an enumerated value such as number of consonants (large, average, small) or how involved tone is (not at all, a little, a lot). a correlation is when the values of two or more features in two or more languages have a statistically significant relationship with each other, such as if a feature having a particular value almost always means another feature will have a particular value.
This visualization and associated tools allow the exploration of features researched for the languages of the world, with the strength of correlations between these features shown as a force directed graph and matrix.
To run this application locally, simply pull the entire /application directory and run index.html. The data is preloaded into preprocessed_data.js, but the visualization will absolutely work with any other data loaded into that file if kept in the appropriate format!
A live demo can be found here.
Our final proposal can be found here.
Our final video presentation can be found here.