Inspiration
Since we are a group of students, we have been seeing our peers losing motivation for school due to un-engaging lectures. Young adults are more stressed due to added responsibilities of taking care of younger siblings at home and supporting their family financially. Thus, students who have more responsibilities due to the pandemic miss a lot of classes, and they clearly don't have a lot of time to re-watch a one-hour lecture that they've missed.
This was me, during the earlier months of the pandemic. By having to work extra hours due to the financial impacts of the Coronavirus, alongside the inaccessibility of the internet when being outside, I missed the most important classes to finish my requirements for my degree.
That's where the inspiration of this project came from. I personally know people in different fields facing the same issue, and with my team, I wanted to help them out.
What it does
By taking an audio file as the input, we use the Google Cloud API's function to turn the audio into text. We then analyze that text to determine the main topics by word frequency which we display to the user. We then display all sentences containing words with the highest frequency to the user.
How we built it
First, we laid out the wireframe on Figma; after further discussion, the dev team went on working on the backend while the design team worked on the high-fidelity prototype.
After the high-fidelity prototype was handed off to the developers, the dev team then built the frontend aspect of our product to allow the user to select audios which they want to condense.
Challenges we ran into
While building the backend, we ran into numerous bugs when developing algorithms to detect the main topics of the audio converted text. After resolving that issue, we had to figure out how to use the Google Cloud API to convert the audio files into text to pass into our processing algorithms. Finally, we had to find a way to connect our website to our backend containing our text-processing algorithms.
Accomplishments that we're proud of
Our team figured out to convert speech to text and to display the output of our text processing algorithms to the user. Our team is also proud of creating a website that displays what our product does, acting as a portfolio of our product.
What we learned
We learned how to utilize APIs, develop algorithms to process our text using patterns, debugging our code while learning under a time limit with new teammates.
What's next for Xalta
- Integration with Zoom and Canvas for a more seamless user experience
- A desktop/mobile native app for stability
Log in or sign up for Devpost to join the conversation.