The problem Safer Instagram solves

The aim of our project is to create a website and a mobile application for the judicious and safe use of Instagram. We have developed a deep learning model for predicting if a user is fake or not. The major focus of our project is on creating awareness on how to handle social media attacks and also to share data about how people survived the attacks, along with relevant statistical data.

Target: There are a lot of fake users on Instagram. They misuse the platform and many innocent people, especially teenagers fall into their traps. In extreme cases, they might lead to them resorting to unethical practices like suicides after facing social media attacks. Our project aims to make Instagram a safer platform and guide users to use the platform in the best possible ethical manner. The government offers support and help to people who face social media attacks. This is often not known by many. We provide the helpline numbers to connect and also the necessary websites so people won’t feel helpless during such an attack. We also provide step by step approach on how to handle these attacks.

Target age group: Can be used by any Instagram users, but the main target is teenagers who themselves cannot distinguish between what is right or wrong on a social media platform.

Implementation of Deep Learning Model: We have used Tensorflow to create a sequential model with 3 Dense layers and dropouts(to prevent overfitting). We have used adam optimizer and ran the model through a set of 85 epochs.

Prediction: By entering details of a user( followers, following, post, if profile picture is there, etc), a user can predict if the suspected unknown account user is fake or not. Thereby, avoiding contact with the unknown.

A 14-year-old girl in West Bengal committed suicide due to the circulation of her morphed photographs on social media by a fake person. The motive of us is to avoid such events in the future and make social media, a better place for people, especially teenagers.

Challenges we ran into

Future Enhancements

1. We have used flutter web-view for creating the flutter application. We will upgrade from this to a custom application and eliminate the need for the website in the flutter application
2. We are planning to scrape Instagram and get the required details instead of prompting the user the input many of them.
3. We will host our website on Heroku