CarCents

Making cents make sense

Comment

🚗 CarCents — Making Cents Make Sense

💡 Inspiration

Buying a car is one of the biggest financial decisions people make — yet it’s often confusing and overwhelming. Between credit scores, APRs, down payments, and loan terms, many buyers struggle to understand how their financial choices actually affect their monthly costs.

We wanted to create something that helps people make sense of car financing. That’s how CarCents was born — an app that empowers users to compare Toyota models, receive personalized financing recommendations, and interact with an AI chatbot that guides them toward smarter car-buying decisions.

Our motto: “Making cents make sense.”

🚀 What It Does

CarCents helps users:

  • 🧮 Compare Toyota cars side by side

  • 💵 Get a recommended down payment and APR based on their credit score and budget

  • 🔄 Adjust down payment and loan term to instantly see updated APR predictions powered by an ML model

  • 🤖 Chat with an AI assistant (RAG chatbot) that recommends cars tailored to user preferences

  • 🔍 Filter and explore real car listings online based on chatbot suggestions

In short, CarCents makes it easy to understand how your credit, budget, and decisions impact the total cost of owning a car.

🧠 How We Built It

CarCents is a full-stack web app powered by machine learning and conversational AI.

🧩 Tech Stack

  • Frontend: React.js

  • Backend: Flask (Python)

  • Machine Learning: scikit-learn, NumPy, pandas

  • Chatbot: Retrieval-Augmented Generation (RAG) pipeline using a language model and car dataset

  • Deployment: Hosted locally for the hackathon

⚙️ Machine Learning Models

We trained two predictive models:

  1. APR=f(credit score,loan term,price,vehicle age,down payment rate)APR=f(credit score,loan term,price,vehicle age,down payment rate)
*   Credit score

*   Car price

*   Loan term

*   Vehicle age

*   Down payment rate
  1. If P(default)≥0.5⇒“Increase Down Payment”If P(default)≥0.5⇒“Increase Down Payment”Else ⇒“Good Standing”Else ⇒“Good Standing”

Both models were trained using scikit-learn and integrated with Flask for real-time predictions via RESTful APIs.

💬 Chatbot Functionality

We built a Retrieval-Augmented Generation (RAG) chatbot that lets users:

  • Ask questions like “What’s the best Toyota SUV under $30k?”

  • Get recommendations that match their preferences

  • Automatically filter car results on the frontend based on AI responses

The chatbot connects natural language queries to structured data, bridging human intuition and real-world car options.

💻 Frontend Experience

Our React frontend provides:

  • Dynamic forms for entering credit score and budget

  • Real-time APR updates when users adjust loan parameters

  • A clean dashboard to compare models visually

  • AI-powered recommendations that can instantly filter search results

We focused on making the UI intuitive and interactive — financial literacy shouldn’t feel intimidating.

🧩 What We Learned

  • How to train and deploy machine learning models in a Flask backend

  • How credit score and down payment influence financial models and APR predictions

  • Building a RAG chatbot for intelligent, context-aware recommendations

  • Managing data flow between React and Flask for smooth user experience

🚧 Challenges We Ran Into

  • Model tuning: Getting realistic APR and risk predictions required multiple iterations and hyperparameter tuning

  • Frontend-backend integration: Handling async calls and real-time updates efficiently

  • Chatbot integration: Translating chatbot output into actionable filters on the car list

  • Balancing accuracy and clarity: Presenting financial data in a way that’s both precise and easy to understand

🏁 Accomplishments We’re Proud Of

  • Building an end-to-end web app that combines AI, ML, and finance

  • Deploying working ML models that provide meaningful financial insights

  • Creating a chatbot experience that feels genuinely helpful — not just gimmicky

  • Designing a platform that could genuinely help people make smarter car-buying decisions

🔮 What’s Next for CarCents

  • Expand beyond Toyota to include all major car brands

  • Integrate real dealership APIs for live listings

  • Add voice-based AI interactions for accessibility

  • Improve ML models using larger datasets and real-world loan data

👥 Team & Roles

  • Machine Learning & Backend: Built Flask API endpoints, trained models

  • Frontend Development: Designed React UI and implemented dynamic car comparisons

  • Chatbot Development: Created the RAG pipeline for personalized car recommendations

CarCents bridges the gap between financial awareness and car shopping, helping users understand how every financial decision affects what they drive — and how much they pay.

Making cents make sense. 💰

Share this project:

Updates