This project was really big, so I'll outline some of the main
- Portal: Login for Professors. Professors could start
and stop lectures
- Transcription : Real time speech to text
transcription of what the Professor is saying in
class that all students in the
class can view.
Phrase Analysis: Entity Analysis for the phrases
before they are transcribed. Articles that
are found and then the phrase becomes a hyperlink in the
Assisted Note taking: Professors can pass in a google
slides file and the text is written
an edittable text file. This makes it so students do not
have to rush to write the bullet points on
and can instead focus on clarifications.
The Digest: At the end of each lecture session, a
digest is composed of 3 main parts, a
summary of the transcription of the lecture, the notes that
the student wrote in the notes tab, and
the important phrases along with a 2 sentence summary for
each one. This makes it optimal for
before a test or just review in general.
How does Megafind work?
This project was fairly massive (for a hackathon) in terms of
the amount of features it had and the time it was built in.
To get the initial set up, we had to get an idea of what we
wanted to do. We started set up by getting a
of the Google Speech to Text API and a summarizer API from
We then used the Google Natural Language API to create an Entity
Extractor that would find relevant
phrases in our text during the real time transcription process.
Following this, we used the Google
Search API to
find a link for all the relevant phrases. The relevant phrases
would then be turned into hyperlinks when
Using the Slides API from Google, we were able to parse the
provided powerpoint from teachers and write
it into an edittable text tab.
To create our digest that is sent to the students at the end of
the lecture session, we gathered the
- A summary of the transcript
- Notes that the student wrote during the live lecture
- The key words and a short 2 sentence summary about them
We then wrote this data into a textfile and sent it. We gathered the
summary by using the Intellexer API,
notes were gathered from the notes file, and the key words were also
go from storage; however, we had to use
Google search API to find 2 sentences about them. We stored the data
about which students to email
through Mlab and we actually sent the email by using the Sparkpost