Speech to Text proof of concept
Google Cloud API
We were inspired to solve the problem that you can't type as fast as you think and even if you do type fast, it is hard to revisit your thought processes as they occurred.
What it does
Think Freely allows you to go back in time and visually see exactly what you were thinking while you were writing. No longer will you be confused by what you meant when you wrote that sentence 3 paragraphs ago.
How we built it
Think Freely was built using node.js and custom data visualizations on top of Google Cloud Natural Language API and Google Cloud Speech API. We prototyped a Angular 4 application and a AngularJS application because there is a full stack project generator that allowed us to get server side code running much faster and begin rapid building. However, we ended up moving to a pure node.js frontend using some express and jade.
Challenges we ran into
- Nobody had experience integrating multiple pieces of dynamic JS with a frontend application, so we spent a lot of our time struggling with this
- We had to develop a unique data structure for storing and accessing English language sentence relationships and dependencies
- We had to develop our own text editor to allow precise manipulation of individual words
- We had to develop our own node based data visualization that allows for user interaction and future version control intergration
Accomplishments that we're proud of
We are able to track the thoughts of a writer as they write and allow them to revisit their past thought process
What we learned
- We learned a lot about English syntactical structure
- We learned how to build efficient and specialized data structures for NLP
What's next for Think Freely
- Ability to share a specific Think Freely doc
- Ability to upload documents for analysis
- Move to a different front end platform for better control over visual design aspects
- Utilize our ability to detect specific pieces of Art/Music/People to display images of said Art/Music/People
- ML to better hone the visualizations and processing of text for a specific profile