Safe Tweets - CatHacks-Submission-2020

The application is for women safety which analyses the tweets of entered username and display the suspect and their offensive tweets.

Comment 2

Tweets-Molester-Finder

Today, women's safety is one of the important concerns in every country and mainly online safety. The offensive messages, pictures, tweets are not only against women but the whole society. Some anti-nationalists will always try to push a country down and those people should be caught. For this, I've made an analyzer that analyzes tweets from tweets just by entering a suspect by their username, or random tweets can be searched by entering a hashtag for a particular word. Whether the tweets are in English, Hindi, or Hinglish, it will analyze the tweets and mark the offensive tweets and give the result. The model is trained in such a way that it can analyze even images from the tweets ( for eg: fir testing purposes we have entered the username of a comedian - Kunal Kamra, {twitter username: kunalkamra88 }, an offensive image link is displayed in the result. ) Further, it gives all the information present on their twitter profile. We can also add a list of usernames simultaneously to give all the results in common. We can also add the suspect in a suspect list present to search their tweets again. I've also made some APIs in Django for analyzing username and hashtags and also for the suspect list which can be used in other apps too and hence can share the same database. The model is not 100% accurate ( the accuracy of the model is more than 93%, trained on 20,000 datasets ), still, it is a push towards a better world. Thanks :)

GitHub repo: https://github.com/raghavdhingra/CatHacks-Submission-2020 Development Server link: https://cathacks2020.herokuapp.com/ I've connected a domain with it: https://safetweets.tech

Share this project:

Updates