Positive Panda - Hack 2021
- 1. P O S I T I V E P A N D A A B R O W S E R E X T E N S I O N T O H E L P Y O U R E M O V E N E G A T I V I T Y F R O M Y O U R D A I L Y N E W S A N D I N D U L G E I N T H E P O S I T I V I T Y O F T H E W O R L D
- 2. W H Y P O S I T I V E P A N D A ?
- 3. Why Positive Panda? • Our AI powered solution “Positive Panda” can tell if the news articles on Google News are positive or negative • It also finds the salient keywords from the news article to display to the users.
- 4. Daily News served right to you • Positive news served daily to your doorstep. • Users can click our extension icon on the top to view the top positive news from around the world. • This selection of news is updated daily. • Users can choose between categories Why Positive Panda?
- 5. All the Control with All the Ease Why Positive Panda? We provide you the ability to proceed with care. The users are not only provided with the notable keywords but also a rating so that they may assess whether they want to view the news article. We ensure user privacy is maintained by not sending user data over a server
- 6. T A R G E T A U D I E N C E
- 7. All Demographics • We aim to target people of all demographics, as our tool allows people to enjoy a seamless and positive experience while browsing the news • Our product aims to spread love by limiting violent and negative content while also suggesting the positive ones !
- 8. M A R K E T A N A L Y S I S
- 9. 1. Soothe • An extension to blur out articles that have homophobic or trans-phobic terms • Only works if the article specifically has specific terms that are considered trans-phobic or homophobic • Does not flag articles if they are violent or negative but do not explicitly contain such words. Our solution can flag these articles as well since it is AI driven Market Analysis
- 10. 2. Social Street Smart • More focus on detection of fake news • Our solution is unique as it aims to completely filter out violent and negative news instead of filtering fake news Market Analysis
- 11. T E C H N I C A L S L I D E S F R O M H E R E , W E W I L L G O U N D E R T H E H O O D T O D E S C R I B E T H E T E C H N I C A L A S P E C T S O F O U R S O L U T I O N
- 12. W H Y G C P A N D M O N G O D B ?
- 13. Google Cloud Platform UsingGCP'sKubernetesEngine,GKEwe havethe abilitytoindependentlyand automaticallyscaleeach module with popularityanddemand. Scale UsingGCP'sContainerRegistry,GCRwe don'thaveto worryaboutimagehosting andsecurity. Security
- 14. MongoDB Atlas MongoDB scales with our users and popularity because MongoDB’s horizontal, scale-out architecture can support huge volumes of both data and traffic. MongoDB Atlas makes it very easy to have an instance running in all major Cloud Providers,including GCP. It removes the headache of setting up and managing MongoDB in a secure way.
- 15. P I P E L I N E
- 16. Flow of the Application User visits Google News to get their news Theextension sendsthe news articles to the backend server. Theserver comparesfor pre-computed resultsfor the news article in the MongoDB database. If found,the server returnsthe results,else the server queuesthe news article on Rabbit MQ and returnsa <UNK> result. Thequeued articles are sent to the Sentiment Analysis model runningonthe KubernetesEngine on GCP. Thecomputed resultis pushedonto another queuewhich is sent to the Flask server to pushthe resultsto the database. (Change:The model sendsa post requestwith the computed results to the Flask server to add to the database) Once the user extension receives the results,it blursthe relevant news articles. If the responseincludes<UNK>, it blursand displays a <UNK> tag to the user.It then sends another request to the server after a set interval to update the <UNK> articles.
- 17. The Scoring Model The model works by building upon the following libraries: 1. Newspaper3k: This is used for scraping and parsing the content of the articles. 2. Huggingface: The sentiment Analysis model is implementedusing the HuggingFace library.
- 18. H A P P Y S U R F I N G !