Track: Pioneer
Inspiration
As CS students, one of things we struggled with early on was CS2: it was a big jump from CS one and data structures seemed daunting and complex. We were also Calc students studying complex integrals, derivatives, and more. While there were websites to help us calculate these problems, there was never a central location that was able to cover aspects of both math and CS that was user-friendly. Fibonacci aims to solve this problem by providing an easy-to-use app that generates videos you can use to help visualize.
What It Does
Fibonacci is a graph and algorithm visualizer that is the one-stop place for students to visualize complex math or CS topics that they’re struggling with. Using a cutting-edge rendering system, students can input the exact problem they’re struggling with, not just select from a predefined list. While currently it only supports math and CS, it could be expanded to future topics such as economics, physics, and chemistry. This is the main appeal: this could essentially be the Khan Academy of visualizers, used for a variety of subjects.
Technical Overview
Our system architecture consists of four main components: the frontend, backend, visualization engine, and containerization setup.
Frontend (JavaScript, HTML/CSS): Our frontend is built using JavaScript, HTML, and CSS to keep it lightweight If we were to push the project into full production, we would use a framework such as React or Vue The frontend handles: User input for selecting mathematical functions or concepts to visualize Displaying generated visualizations or videos returned from the backend
Backend (Django) The backend, powered by Django, acts as the central hub that connects the frontend with our visualization model. The backend handles: Handling user requests and managing input parameters (function type, range, animation style) Triggering the Manim scripts that generate the mathematical animations Managing storage and retrieval of generated animations or visualization data Exposing API endpoints for the frontend to fetch results Ollama was also used to create the AI-powered insights section with a lightweight LLM.
Visualization Engine (Manim) The core of our project lies in Manim, a powerful open-source Python library created by popular Youtuber 3Blue1Brown for generating mathematical animations. We figured this was a unique and cutting-edge engine for edtech and could revolutionize the way we can look and visualize theoretical concepts in math in CS. When a user inputs an equation or visualization request, Django executes a corresponding Manim script to generate a video or animation that illustrates the concept. These animations are rendered on the backend and then sent back to the frontend for display.
Containerization (Docker) We used Docker to containerize the entire application, ensuring consistency and portability across environments. Our Docker setup includes: A base image with Python and Manim preinstalled A Django container running the backend server A frontend container serving static files and handling user interactions
System Flow The user inputs a mathematical expression or chooses a concept. The frontend sends this request to the Django backend through an API call, the PLACEHOLDER value is assigned to a variable. The backend uses a Python Manim script to render the corresponding animation, with the inputs that we used with the API call. The animation is saved and returned as a video to the frontend. The frontend embeds the animation in the user’s browser.
Challenges We Ran Into
Using Docker, there were repeated discrepancies between the localhost and live server output, the frontend kept disappearing. It turns out container isn’t correctly including the file, so after we added the static file collection and changed the file path, it worked.
Future Improvements
Implement a modern front-end framework (React or Vue) Integrate cloud-based rendering to handle heavier animations Add user accounts and a visualization history Expand to different academic subjects