With one out of every five US citizens having a foreign background, communication within multi-lingual communities has never been so relevant. Conventional solutions are often very inconvenient or expensive, usually involving downloading and installing a mobile application. Foreign speakers need a more lightweight, easy solution to help communicate effectively with other members within their community.
Our solution is a service that operates as a link management platform. But unlike conventional shortening services, the aim of our platform is to more efficiently aggregate information to people of different backgrounds regardless of the content we are sharing and tear down the language barrier completely.
What it does
On top of maintaining links, we parse through the content and use Watson's cloud services to accurately translate and rebuild the content in a cohesive fashion. Unlike Google Translate services that blindly go through the webpage and tries to translate every sentence it sees, our service takes a more logical approach and more accurately translates our content.
How we built it
We created a web platform that can manage links and store translated content to be shared and redistributed. For the web parsing, we used Python's Beautiful Soup library (https://www.crummy.com/software/BeautifulSoup/bs4/doc), and used Django as our web framework interacting with a Postgres Database.
Challenges we ran into
Accomplishments that we're proud of
We are proud to have been able to create a working product that is able to interact with real webpages. We were not even sure if this was going to work and it took a lot of ingenuity and creativity from our team to be able to get this done.
What we learned
We learned that Watson's API's are very rich in what they are capable of accomplishing as well as the feasibility of their use. We will definitely try and utilize their API's again in the future!
What's next for Smart Bitly
Increase the speed of translations. It takes a while for translating pages with a lot of web elements as our algorithm goes through the source code explicitly to translate. We will employ parallel computing as well as multithreading to try and increase the speed of our platform