The political world is filled with legal jargon that adds a lot of filler in legal texts, such as regulations, contracts, bills and more. Recently in the United States a 5593 page COVID Relief Bill has been sent to be voted on by Congress. Every citizen has been affected by COVID and the government’s response, yet not every citizen has the educational background or time to analyze exactly what the COVID Relief bill was about and how it would serve them.
Every citizen should be able to understand the legal information in a bill or at least the summary of it. Especially bills that affect marginalized and oppressed demographics.
What it does
Using the search bar
- Features of this: Search bar via bill title or some parameter
- Technology: searching a firebase db to check if bill exists
Bills show up in table
- Features: Table that displays bill info
- Technology: Info displayed on React front end home page
User could also upload an URL where
- Features: Modal popup that
- Technology: server through which data is scraped and passed onto the NLP Model. After processing the summarized results, it is pushed onto the Firestore database and the info is displayed on the bill page.
User clicks bill to view if they find the desired bill
- User gets redirected to bill page
- Feature: Button to redirect
- Technology: Firebase db has to be queried to retrieve the summarized bill, which has been processed by Python NLP Model backend
Challenges we ran into
Extracting the information accurately was a challenge because we do not have backgrounds in law or politics. So we had to learn a lot about how bills are structured and the format of legal information inside bills.
Accomplishments that we're proud of
We are proud of coming up with a solution in 48 hours. We are really happy with how the team coordinated and collaborated during this hackathon
What we learned
We learned how to use some new technologies and we learned a lot about how legal bills are structured. We also gained some good experience in developing an efficient database architecture.
What's next for BSBill
- The Python NLP library used is still actively being improved. In the future Information extracted from the legal text will become more accurate as more text analytic functions will be added.
- This project can be extended to various other legal text processing applications outside of political bills (such as employment contracts), which the BSBills team is currently planning to work on.
- Context/Theme Identification technique can be applied in order to provide the user with a summary of what the bill is about. As right now we are extracting info, but a summary would be helpful