Project for Koding.com #hackathon
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
elasticsearch/configs
eulathingy
.gitignore
LICENSE
README.md
logo.png
requirements.txt
ss-app.png
ss-landing.png
ss-results.png

README.md

EULAThingy

logo

About This Project

TP;NS was founded in 2014 for the world's first virtual hackthon hosted by Koding.com. EULA Thingy is a Django-based web application which combines the two powerful elements of crowdsourced information and machine learning. We hope this will provide the online community with a way to quickly summarize legal resources.

What did we actually manage to achieve?

Well, the hope was to create a system where it could locate similar entries and automatically predict the rating that should be applied. However, you soon realise that 48 hours isn't a massive amount of time and when you're trying to do things you've never done before, it gets a lot smaller - And EVEN smaller with two people!

We managed to get the system to pull out the sentences and then allow users to assign a rating, which we then use highlight individual sentences. We're pleased with what we've done and we're going to work on it a bit to both improve our skillset and add some of things we missed.

It was a great experience, thanks a lot to Koding for hosting it!

Screenshots

Landing page

landing

Interactive legal documents with term severity highlighting

app

Document results based on user ratings.

results

Packages

  • NGINX
  • Data Dog
  • Google Analytics
  • Django
  • Python
  • NLTK
  • POstgres