TAMU Datathon First Place Winner! 🎉🎉🎉
Created by Akilan Manivannan, Nicholas Kasman, and Zeeshan Chatur at Texas A&M University.
FishType is a project that can capture and analyze inherent patterns from the way that people type. Formally known as Keystroke Dynamics, our research aims to take the same principles of gait analysis and apply them to typing. Our project captures the specific timing data from each keystroke that you type and uses that to determine your identity. We hope that this can be used as another form of authentication to reinforce existing security systems.
Youtube Presentation: https://youtu.be/a60TJ6N8p7I
The public directory contains the code we used to collect and upload the keystroke data. The front end is available at https://fishtype-3f3cb.web.app/.
The analysis directory contains the Google Colab file along with our official data analysis write up that proves the valididity of the theory that Keystrokes can be used to identify people and even has a model that can classify our dataset with up to 88% accuracy!.
Colab Notebook for Data Analysis: https://colab.research.google.com/drive/1gypNJ5vnazcBVmV-uc6SD3tj2WjU2otD?usp=sharing
While we have plans to upload the dataset to Kaggle or make a link available on our site, we have not done so yet. If you are a researcher and are interested in using our data, please contact Akilan Manivannan ([email protected]) or Nicholas Kasman ([email protected]).