Generic medical dosages don't account for individual differences in genetic makeup, they only account for the "average human. But there is no "average human". Medi-Twin solves this by creating a personalized digital avatar. By calculating metabolic clearance rates in real-time, I turn invisible body chemistry into an interactive 3D experience.
Medi-Twin creates a personalized 3D avatar that simulates how the body processes substances like caffeine or medicine. By entering a weight and metabolic rate, the app solves biokinetic equations in real-time to show the "digital twin's" current concentration levels and estimated recovery time.
I developed a Python FastAPI backend to solve first-order kinetic ODEs using SciPy. For the frontend, I used React and Three.js to create a responsive dashboard and a 3D mesh that physically distorts and changes color based on the simulated data.
The most difficult hurdle was the logic bridge: connecting the output of a complex mathematical ODE (Ordinary Differential Equation) to the physical properties of the 3D sphere. I had to normalize biological data into values that could drive real-time mesh distortion and emissive lighting without lag.
I am incredibly proud of how the 3D Digital Twin turned out. Seeing a mathematical concentration value successfully trigger a physical "toxic" state, where the sphere turns red and vibrates, perfectly captures the biological stress I wanted to visualize.
What I learned I gained deep experience in Full-Stack Data Flow, specifically how to pass time-series data from a scientific Python engine to a high-performance JavaScript visualizer. I also learned how to simplify complex pharmacokinetic concepts into an intuitive user experience.
What's next for Medi-Twin Expanding the simulation to account for age and caloric intake to increase accuracy. I also plan to add a library of different drugs and refine the UI into a dedicated educational tool for pharmacology students and health-conscious users.
Log in or sign up for Devpost to join the conversation.