We were originally inspired by the recent events in politics in regards to our Immigration policy and the strong response that came as a result. We saw that people, especially lawyers who work with immigrants regularly, are incredibly passionate. However, increasing exposure in recent times have caused overburden to many individual lawyers as well as organizations such as the ACLU, who already work tirelessly on this cause.

What it does is an application that helps keep immigration lawyers on track by prioritizing their work for them based on the severity of the situation of the client. It does this by listening in on the calls that clients make to the office then uses natural language processing to gather data about the conversation. This data would later be used to build client profiles for the lawyers to look back on as well give suggestions to the lawyer on resources that they can provide their client based on their conversation.

How we built it listens in on the call using Nexmo's voice over websocket API. It then converts the speech output to text using IBM Watson's speech to text service. It does performs Natural Language Processing using both IBM Watson NLP service as well as Python NLTK library. It then takes the data and build profiles based on decision trees that we've built out based on our research to help categorize the clients. Lastly, it uses the profiles that was just built to search for relevant resources.

Challenges we ran into

Learning new APIs in a short amount of time is difficult. NLP is also always hard.

Accomplishments that we're proud of

It works.

What I learned

We learned how to leverage the Nexmo and IBM Watson APIs.

What's next for

Support for more types of laws or industries. Also support for multiple languages.

Share this project: