Inspiration can record public opinions on a subject and the Federal Register has the option to take public comments on legislation. Both of these platforms have their own issues. opinions are generally invalidated by the fact that users aren't verified. The Federal Register is a less accessible platform that not many users know about and even fewer are interested in reading long legal documents. We want to make the legislation more accessible to the public by summarizing it and capturing real user opinions.

What it does

MotherNature is fed pending action documents relevant to nearby national parks from the Federal Register. Our summarizer uses machine learning to summarize documents on our database periodically. Users can find documents that are most relevant to their area using geolocation and sign their approval or disapproval using the DocuSign embedded signing API.

Challenges we ran into

  • Docusign Identity Verification APIs (IDV and KBV)
    There are a few issues we ran into with “Recipients View API” vs “Envelope API”. The latter was significantly easy to use whereas the former provided with a much better UX due to the embedded nature of it. We stuck with the recipients' view API. The ecosystem of the supported languages was not in our favor, we started with Python for the eSign API but later realized that nodeJS was better supported for IDV or KBV.
  • Federal Registry Integration
    Our original plan was to fine-tune our text-summarization model using the summaries for the past legislation in the federal registry. The site did not have any APIs and was not structured well for scraping. So, we decided to stick with the Harvard-NLP dataset, which seemed to perform reasonably well.

Accomplishments that we're proud of

  • We were able to nail down our idea pretty early on which let us focus on providing the best possible UX.-
  • We have a fully running app with all material design elements and integrated with embedded DocuSign signatures.
  • Scalable backend running on Google Cloud AppEngine.

What we learned

  • A lot about the efforts that are in progress for the Wilderness Society.
  • Talking to many of the DocuSign engineers helped us learn the various challenges that need to be tackled for identity and signature verification all the way from on-device machine learning as well integrating with public data to come up with KBV challenges.
  • Developing and deploying with Google Cloud and Firebase was extremely easy.

What's next for Mother Nature

The future of Mother Nature consists of first finding a workaround to implement the IDV / KBV for the application. Another feature we plan on implementing would be a visual graph displaying the trend of voters on the admin panel of the applications.

Share this project: