Ripple_Engine

Dashboard
Landing Page

Inspiration

I got the idea for this project after noticing how many developed countries carefully plan and simulate cities before construction. In India, especially in Tier 2 and Tier 3 cities, I often see newly built roads being dug up again within months to lay pipelines, drainage, or cables because different departments work separately.

This leads to inconvenience for citizens and wastes public money, time, labour, and resources. I wanted to build a solution that helps governments simulate the impact of infrastructure decisions before construction begins, enabling smarter planning and reducing avoidable mistakes.

What it does

Users can create a planning scenario, such as building a new road, metro station, shopping mall, or residential area. The platform then simulates the possible ripple effects on traffic, pollution, flood risk, public services, infrastructure demand, and the local economy .

Instead of simply showing predictions, Ripple Engine explains why those outcomes are likely and provides recommendations to reduce potential risks. This enables decision makers to compare different planning options and choose the one with the best long term impact for the city.

How we built it

I engineered the backend using Python and FastAPI. For the core intelligence, I integrated the Google Gemini 2.5 Flash API to process natural language proposals and route them to the correct infrastructure category. Once routed, a custom Monte Carlo engine built with NumPy runs ten thousand permutations to predict urban impact. I also connected live weather and air quality APIs to feed real time data into the simulation. For the frontend, I used plain HTML, JavaScript, and Tailwind CSS to design a responsive brutalist dashboard. Finally, I deployed the frontend on Vercel and containerised the backend using Docker for hosting on Railway

Challenges we ran into

the challenge was figuring out how to bridge a natural language model like Google Gemini with a strict mathematical Monte Carlo engine. Furthermore, writing the Monte Carlo algorithm from scratch to deeply understand the underlying probability mathematics proved to be quite complex and time consuming ( also use ai )

Accomplishments that we're proud of

I am incredibly proud of successfully combining a large language model with a strict mathematical simulation engine. Instead of a basic text wrapper, Gemini acts as a semantic router that triggers hard calculations. Treating city infrastructure like a network where I can patch vulnerabilities before they happen was a huge win

What we learned

I learned how to build and integrate stochastic probability models into web applications. Upgrading the Gemini SDK taught me the importance of reading updated documentation and handling legacy code transitions under strict time constraints.

What's next for Ripple_Engine

My immediate plan is to scale the platform beyond the Lucknow minimum viable product to support dynamic city mapping across India. Adding node system which run and get more data and acheive more accury and the feature i have build make it workiing soon.