PROJECT
SnapSeek
The problem SnapSeek solves
In today’s digital age, YouTube has asserted itself as a vital piece in learning and decision making for users of all generations. NPTEL, MIT OCW, Stanford Online, Tech tutorials, etc. are integral for today’s engineers, and the underlying platform for all of these is YouTube.
The problem that we identified is that, when vast content is amazing as it gives tremendous data in the hands of the user, it is also leading to a problem of plenty - the user has to spend a lot of time browsing content before he finds the piece of information he is looking for. This is where our solution comes in.
This solution aims at reducing the time involved in finding the right piece of information - the pièce de résistance, and quicken decision making, with more meaningful data. We have built a clean and simple solution which does a query-based search of videos for keywords (Educational videos, Online tutorials, News Headlines, TED Talks), and an Entity-based Sentiment Analysis (Product reviews, Comparison Videos).
We have the following features that help us tackle the problems listed above. A Keyword-based search of YouTube videos, which gives a set of timestamps. These timestamps are the locations in the video where the query was discussed. The user can directly jump to that point in the video that interests him. The user can also avail of entity-based sentiment analysis (again based on keyword query), and get a graphical representation of an ad-hoc Sentiment Score (this is particularly helpful when considering review videos and comparison videos).
