We know our hack can be the winner of HackNotts. How can we prove it?

What it does

HackProphecy scrapes 35,000+ hacks metadata from DevPost (sorry!), and feeds these to a machine learning algorithm. It predicts whether any given hack could be winner or not based on some natural language processing and support vector machine algorithm.

Accomplishments that I'm proud of

We can boast an ~94% accuracy.

Share this project:


posted an update

Project is now online at, hosted on OpenShift PaaS cloud.

Still working to get external connections to work on OpenShift, in the meanwhile the following projects should work as demo search terms for the project:

call-serena-button dogfight-elite-vr cronforce-time-tracking-for-jira-dis5by study-1g65v4 updated-issues-basket my-ride rest2ddp teamdocs vais bussi-zg9eqf cycle-control fishhackaton smiley powagri criziz-tyuvo export-recover-exchange-edb-to-pst-converter pray-it-forward-ybna9 sharemail sahara-s3zdw4 lucidchart-onprem-for-confluence-server the-night-cafe-an-immersive-tribute-to-vincent-van-gogh spell-script knowfish sticky-mail delveclassifier dataless-ob3caj the-bouncinators kickback fuji shipply neos-the-universe

Log in or sign up for Devpost to join the conversation.