What is Athena?

Athena is a project originally built for HackRU Spring 2022. It algorithmically searches for the most important sentences in a YouTube video, and summarizes the video using these sentences.

How Does It Work?

Athena's Algorithm Works as Following:

1. Athena turns the mp3/mp4/wav of a YouTube video into a string transcript using a Google Cloud API
2. It then parses through the transcript and calculates the frequency of each word - counting similar words (play, plays, player, playing, etc.) as the same word.
3. It then goes through these words and multiplies the frequency of each word by its correlation (from 0 to 1) to the main topic of the video, assigning this value as the 'score' of each word.
4. Then, Athena calculates which sentences contain the highest scoring words, and outputs the top n for the user to read

See YouTube Video in project description for more info

What is Athena For?

Athena seeks to help students who are on a time-crunch during important study periods. Students are often given small 5 point assignments in classes that are to summarize or make some opinion of a video. Instead of taking 30-45 minutes watching these videos, students can simply use Athena to gain a basic understanding of the video.

What Was Used to Build It?

1. Google Cloud API - Offloaded the speech detection and storage to Google Cloud to increase efficiency
2. NTLK - Used to calculate the correlation between various words in our algorithm. Also used to scrape the transcript for garbage words (the, then, before) and mark them as negligible
3. Python3

Future Additions Coming Soon

1. Updated Framework using HTML
2. Optimization of the scoring algorithm to decrease waiting time
3. Chrome Extension to work directly in your browser

Acheivements

Won 'Best Usage of Google Cloud' in HackRU Spring 2022 @ Rutgers University