Clyde is a chrome extension that uses multiple NLP techniques to analyze sentiments, behavioral traits, emotional traits & IPTC traits for anything on the internet. You can investigate Youtube videos, news articles, tweets, or even subreddits in real-time & get meaningful insights. People often freely express their views and opinions on social media, providing a wealth of information about their thoughts and feelings. Clyde offers a simple way to mine these opinions and monitor reactions on social media by just keeping the relevant parts. The emotions and behavioral traits help us determine whether the people are angry, satisfied, or confused when reacting to a news item. Leveraging's sentiment analysis capabilities, we can establish whether the news item's perception is positive or negative. Using the snipping tool you can screenshot & select specific parts of the web pages across multiple websites to get a detailed report on the topic in question. For example, we want to examine the public sentiment on Bitcoin after China's recent crackdown of the cryptocurrency across Reddit, Twitter, Youtube & news articles. With Clyde, doing this is a piece of cake.

Let us suppose your studying for your history test. You go to to learn more about the topics that will come in the examination. When you come across a pretty long video, you have no way of knowing beforehand whether it covers the test syllabus. Upon clicking the 'Get Video Insights' button, you get a summary of the video along with a word cloud telling you precisely what is inside.

What it does

To get started, you have to pick your topics of interest and a secondary language from Spanish, French, German, or Italian to get the results in your secondary language.


Opinion Mining & Web Content Analysis

Just select the text areas you want to monitor using the snipping tool & click the 'Analyze Images' button. You can take up to 8 screenshots across different websites to get a detailed analysis. You'll get the results back as piecharts & word clouds. Using this feature you can tap into customer insights for your product to make better business. For example, you can screenshot their product reviews to get a detailed analysis.

Youtube Video & News Report Analysis

Did you know that the average youtube video length is about 11.7 minutes? And we all know news reports can be pretty long and monotonous. Your time is precious. How do you use it wisely? Clyde helps you manage your time well by summarising news articles & youtube videos. Tag-based interest matching assists users in filtering out content that doesn't match their preferences. Additionally, for news articles, all the IPTC media topics are listed for your convenience.

Language Translation

For bilingual speakers, you can get the summary & the word cloud in Spanish, French, German, or Italian upon clicking the toggle button on the top right.

How we built it

We built the chrome extension using HTML, CSS & Javascript. The charts have been added using the chart.js library. The backend of the application is written in Python. The REST API is built using Flask. We use AWS texract to get a text from the screenshots using Optical Character Recognition. News articles are scrapped using Beautiful Soup & Regex. Youtube's transcript API is used to get the transcripts for the video. Once the text is extracted in any one of the three ways, it now analyzed using the following's NLP APIs.

  1. Document Analysis
    • Keyphrase Extraction
    • Sentiment Analysis
  2. Document Classification
    • IPTC
    • Emotional Traits
    • Behavioral Traits

All the relevant key phrases & topics are extracted from the youtube subtitles or news articles and compared with the user's preferences, to find out everything in common. These common topics are shared as chips on the extension dashboard.

The key phrases are used to make Word Clouds. Word Cloud is a data visualization technique used for representing text data in which the size of each word indicates its frequency or importance. This gives the user an overview of what the author is trying to say.

The IPTC traits, Emotional traits & Behavioral traits that we get from are shown as piecharts made using the chart.js library


Challenges we ran into

  • Setting up the chrome extension
  • Connecting chrome extension to python backend

Accomplishments that we're proud of

We were able to build a decent chrome extension that brings advanced Natural Language Processing & Opinion Mining capabilities to your browser. Another thing that we are proud of is the fact that we have successfully built an alternative for traditional web scrapping for opinion mining.

What we learned

We learned about how to build chrome extensions. We also learned the basics of NLP & its practical applications using NLP APIs.

What's next for Clyde Chrome Extension

Marketing the extension and adding it to the chrome web store.

Share this project: