OpenGV | Devpost

Inspiration

In today's political climate there is a growing rise in voter apathy. Due to both the overwhelming pace of news and the lack of transparency in the governmental process, this challenge threatens many of the foundations of our democracy.

What it does

Our mobile application allows users to search up politicians to quickly receive their contact information and location data as well as their past voting history and committees they are a member of. You can also search up committees and bills independently and compare multiple politicians with each other.

How we built it

Most of the data collection was done by API calls in several different languages. Image data was scrapped from the internet on ballotpedia.com and all data was stored in the JSON format. The app was created using andriod studio and Java while our own API and server was coded using node.js.

Challenges we ran into

Originally we were looking towards using augmented reality to overlay information over a politician's face using facial recognition. While we were able to gather the reference and training data as well as develop an augmented reality framework, the facial recognition portion proved challenging. We tried 3 different facial recognition libraries including face-app.js, amazon web services computer vision, and a python facial recognition library but they proved either too difficult to implement or required permissions we did not have on our server.

Accomplishments that I'm proud of

We used a variety of API's to gather all of our data and they all had different nuances that made reconciling them difficult. However, we were able to successfully merge image, contact, bill, and other data into a single central source that we could use for our app.

What we learned

Through this project we became more comfortable with SQL, API calls, JSON processing, facial recognition, web scraping, Android studio and more.

What's next for OpenGV

In the future, we are looking to implement the facial recognition aspect that we were not able to at our time at the hackathon and to improve the UI so that it is friendlier to use.