We are inspired by the #MeToo movement during which it took years for victims to come together to seek legal service. We thought that the reason that it took so long was due to the fact that most of the victims mistakenly thought that they were alone and that if there was a service that allows anonymous submission, it would be easier to disclose a traumatic experience.
When you first login into Tether, you are greeted with two options: report an incident or enquire more about Tether. In enquire more about Tether we link counselling services and just an overview of what the app does. When you press report an incident, you are met with a series of form that you can fill: location of the incident and descriptions of the offender. Tether then stores that data anonymously and uses an algorithm to attempt to match you with other users who have filled out similar forms. When the algorithm determines that two users' experiences are similar, they are allowed to then communicate and the app will tell them what aspects of their experiences are similar. Lastly, we leave it up to the user to choose whether or not they want to disclose their experience to the other person; however, all that we do is that we provide the platform that enables their communication.
We built the app in android studio. We first store the user information anonymously, then each variable from the form is quantified and based on similarity, we assign a score to how similar each variable is. Then we do that for the entire form and then we are left with a match probability. We pick the highest match probability to match with the user first.
It was our first android project and our first mobile project in general so we ran into many issues regarding setting up the environment and bringing all the different parts of our project together.
We managed to develop the algorithm successfully, we managed to get the messaging services work, and we manage to design a user-friendly and modern UI.
Next time, we should spend more time at the beginning developing a reasonable scope and plan of action for the project. We were so excited to begin that we immediately jumped into writing code, and this caused us problems later on because we did not understand how to integrate the different parts we built.
First thing on the agenda is to cross-platform with IOS. Then we are looking into applying machine learning for the matching algorithm so that the software is better at recognizing similarity in situations where visual description isn't exactly clear. Lastly, we are looking to instead using verbal description, use new police sketch technologies where an array of faces is presented and the user chooses the one that resembles the most.
Log in or sign up for Devpost to join the conversation.