LII: Under the Hood
On May 12, LII engineers Sylvia Kwakye, Ph.D., and Nic Ceynowa hosted a presentation by the 14 Cornell University Masters of Engineering students they’d supervised this spring as they presented their project work on the Docket Wrench application to LII and Cornell Law Library staff.
LII adopted the Docket Wrench application from the Sunlight Foundation when it closed its software development operation last fall. Originally developed by software engineer Andrew Pendleton in 2012, Docket Wrench is designed to help users explore public participation in the rulemaking process. It supports exploration by rulemaking docket, agency, commenting company or organization, and the language of the comments themselves. It is a sprawling application with many moving parts, and when LII adopted it, it had not been running for two years.
On the infrastructure team, Mahak Garg served as project manager and, along with Mutahir Kazmi, focused on updating and supporting infrastructure for the application. They worked on updating the software and creating a portable version of the application for other teams to use for development.
The search team, Gaurav Keswani, Soorya Pillai, Ayswarya Ravichandran, Sheethal Shreedhara, and Vinayaka Suryanarayana, ensured that data made its way into, and could be correctly retrieved from, the search engine. This work included setting up and maintaining automated testing to ensure that the software would continue to function correctly after each enhancement was made.
The entities team, Shweta Shrivastava, Vikas Nelamangala, and Saarthak Chandra, ensured that the software could detect and extract the names of corporations and organizations submitting comments in the rulemaking process. Because the data on which Docket Wrench originally relied was no longer available, they researched, found a new data source, and altered the software to make use of it. (Special thanks to Jacob Hileman at the Center for Responsive Politics for his help with the Open Secrets API.)
Deekshith Belchappada, Monisha Chandrashekar, and Anusha Morappanavar, evaluated alternate techniques for computing document similarity, which enables users to find clusters of similar comments and see which language from a particular comment is unique. And Khaleel R prototyped the use of Apache Spark to detect and mark legal citations and legislation names from within the documents.
So, where is it?
The good news is that after a semester of extremely hard work, “Team Docket” has Docket Wrench up and running again. But we need to ingest a great deal more data and test to make sure that the application can run once we’ve done so. This will take a while. As soon as the students have completed their final project submission, though, we’ll be starting a private beta in which our collaborators can nominate dockets, explore the service, and propose features. Please join us!