The problem YouTube Comment Insights solves

The like/dislike counter for a YouTube Video can't describe the entire quality of a video. With YouTube taking out the dislike count, it becomes hard for people to judge the quality of the video.
So, this product is focused on the following Use Case Categories:

1. YouTube Creators
2. Brands Collaborating with Creators (Marketing Aspect)
3. Quality Educational Content Identification

• Approximately 500 hours of video are uploaded to YouTube every minute worldwide. For some results of search, we get lots of video suggestions and all of them have similar tags and descriptions, this will create a lot of confusion and chaos in consumers for extracting the helpful videos and channels with actual content.
• Also, for video creators, it becomes very tedious to go through all the comments and find the interest of their audience.

Solution:
To solve the above problems, we have Sentiment Analysis for our help. Sentiment analysis is a machine learning technique known for opinion mining that is used in generating labelled reviews from comments of users.
The idea is to use the experience of the previous consumers' audience for analyzing the quality of video. Comments will be collected from YouTube using YouTube API. The analysis of comments will be done for specific videos. By using the customer review system, the assessment of the video can be done and finally the customer can decide whether to consume it or not.

Implementation:
With the sentiment analysis provided by Google Cloud NLP API and statistical data extracted by Natural Language Pre-processing of comments on the edge, this solution can scale as the traffic changes with the help of Cloud Functions. The frontend is developed by using React JS making the application fully serverless leading to minimum infrastructure cost.
Key Features:

1. Multilingual Comment Support (15+ Languages)
2. Quick Results (< 1 min)
3. State-of-the-art Prediction Model

Challenges we ran into

The main challenge was to decide the architectural flow for this project and to come up with a product in a very small amount of time. So we checked the availability of data sources and with proper discussion and planning came up with the architectural flow provided below. The architectural flow and distribution of work were according to proper software development paradigms hence leading to a great product in a small period of time.