Autumn

Turn actions into futures

Comment

AUTUMN - Environmental Simulator

Inspiration

Environmental degradation occurs gradually, making its long-term effects difficult for people to perceive. Many individuals fail to understand how everyday actions influence ecosystems over decades. AUTUMN was inspired by the need to make these slow, invisible changes visible and understandable. The project aims to help users explore how forests, oceans, and cities evolve over time under different environmental and human-driven conditions.

What it does

AUTUMN is an AI-powered environmental simulator that models the long-term behavior of three systems: forests, oceans, and urban air quality. Users can modify environmental and human-related parameters such as deforestation, pollution, industrial growth, and waste generation. The system simulates how these factors affect ecological health over time using mathematical models. The results are displayed using static bar graphs and time-based trends, along with AI-generated explanations.

How we built it

Each module was designed as a time-step simulation where the system state updates annually based on influencing variables:

[\S_{t+1} = f(S_t, X_t)]

Here, (S_t) represents the system state at time (t), and (X_t) represents the influencing factors.

The logic was implemented using Python. Data visualization was done using static charts to keep the MVP simple and accessible. The Gemini API was integrated to generate natural language explanations that describe the simulation outcomes in a user-friendly manner. The MVP focuses on transparency, clarity, and educational value rather than complex real-time modeling.

Challenges we ran into

One of the main challenges was designing models that were simple enough for an MVP but still scientifically meaningful. Environmental systems are complex, and reducing them to understandable formulas required careful abstraction. Another challenge was normalizing variables with different units and scales. Integrating AI-generated explanations that accurately reflected numerical results was also a significant technical hurdle.

Accomplishments that we're proud of

We successfully built a multi-module simulator that connects mathematical modeling with AI-driven explanations. The system clearly shows how environmental factors interact over time. We created transparent formulas, interpretable outputs, and an educational interface that makes environmental change understandable to non-experts.

What we learned

We learned that explainability is critical when dealing with scientific simulations. Users must understand not only the result, but also why it occurs. We also gained experience in time-series modeling, system abstraction, and using AI as a communication tool rather than just a prediction engine.

What's next for Autumn

Future versions of AUTUMN will include geographic visualizations, real-world data integration, more advanced machine learning models, and policy simulation tools. The long-term goal is to evolve AUTUMN into a decision-support platform that helps users, educators, and policymakers explore the environmental impact of their choices.

Built With

Share this project:

Updates