Detection of Child Predators/Cyber Harassers on Social Media

Introduction

The tool intends to identify the cyber predators/child harassers on social media with a malevolent intend. The posts, comments and followers on social media are subjected to analysis using Artificial Intelligence, Machine Learning with IGPL and NSFW(Not Safe For Work) to categorize the offensive contents.

This system is capable of analyzing all social media platforms like Instagram, Twitter, Facebook, etc., and other outlets seeking the same suspect. If the suspect doesn’t have the same user ID on different platforms, then Reverse Image Searching is done to identify the suspect. A set of user_id is used as a key to grab their personal information and their post information(Post ID, Comments, Timestamp, location, Captions) from multiple social platforms using ​ OSINT(Open Source INTelligence) and Beautifulsoup Python Package.

Demo Link - Link to Code

Analysis

The above data of various posts are subjected to analyze malevolent contents using Machine Learning and Pandas Python library. Based on the statistical analysis, suspects are categorized based on their behavior(also Polite harassment). The users whose suspect level is greater than the threshold value will be scrutinized and monitored for further analysis. The suspected user’s post information(media like Image, Audio, and Video) is retrieved and analyzed using the ​ IGPL Python package, ​ Urllib and ​ Artificial Intelligence with ​ NSFW (Not Safe For Work) library to make them fall under the category ‘suspects/predators’. Finally, the Child grooming patterns followers and statistical results that are generated are analyzed and the concerned person is classified as predator and reported to the law enforcement authorities

Conclusion

The suspected user’s post information(media like Image, Audio, and Video) is retrieved and analyzed using the ​ IGPL Python package, ​ Urllib and ​ Artificial Intelligence with ​ NSFW (Not Safe For Work) library to make them fall under the category suspects/predators. Finally, the Child grooming patterns followers and statistical results that are generated are analyzed and the concerned person is classified as predator and can you automatically reported to Gmail address which has been stored in the server.