Inspiration

Inspired by a generation of content creators, lifestyle vloggers, and role models on Youtube, Twitter, and other social media platforms, we built the LRT -Linguistics Relativity Transfer - Extension. The principle of linguistic relativity holds that the structure of a language affects its speakers' worldview or cognition. It is evident that YouTubers have the power to entertain and influence decision making for their audience across the globe.

What it does

We used the voice of influencers on Youtube provided by the transcripts of their videos alongside tweets from Twitter to conduct semantic analysis on the products they love and share with their viewers. We gather these metrics and make it available for our users to with one click of the button in while you are on Youtube with our chrome extension. You will get data pertaining to the influencer you are watching as well as information relating to the products they are talking about without obstruction to your viewership.

How we built it

Our application is largely being powered by Google services leveraging Firebase for Database, Storage, and hosting our Node.js server. We used the neural net in SCIKIT for Machine Learning. Front End is built with standard web-dev technologies in the form of an extension designed to provide optimal customer experience.

Challenges we ran into

Some of the biggest challenges we ran into definitely came as a result of limited/incomplete documentation of Youtube Data API. For example, we tested sample code on the API site's page that Youtube provides with the relevant supporting files but it didn't run properly. There also wasn't a workaround online and the online general consensus was that it would be fixed in a later update. We didn't accept this answer and after much perseverance, we found the solution buried in one of the included modules by Youtube.

We received a lot of pushback in this area but we pushed through each problem carefully. We also were challenged by how limited the data was when we finally reached the endpoint of the API but manage to get insightful data for our users.

Accomplishments that we're proud of

  1. Pushing through the problems with Youtube's Data API and not accepting defeat despite what sites like Stack Overflow and GitHub users were suggesting.
  2. Helping make up for each other's weak points, learning, and laughing through those difficult problems we encountered. We also had the honor of working with a first-time hacker in our group who reminded us how important qualities like curiosity and determination are.
  3. Making it through the night!

What we learned

We learned how to interact with/further our understanding of various technologies such as Google Cloud Platform tools such as Firebase Database and Storage, SCIKIT, and Sentiment Analysis (more specifically with Twitter API). Aside from that, we realized how rusty our web-dev skills had become and spent the weekend of learning to polish and hone our skills especially with solving backend problems.

What's next for LRT - Linguistic Relativity Transfer - Extension

Short Answer? TBD. Long Answer? Well, we would have loved to build our own custom API (perhaps with Ruby on Rails) with more capability than what the Youtube Data API could provide us with. In a 24 hour stretch with time ticking away, the realization came to us when the possibility of deploying our own API was no longer possible. We would also have approached the UI differently more like a dark yet translucent layover with the visualization of data in the surrounding edges of the Youtube Video.

Share this project:
×

Updates