SpotWise

Interactive Map Diagram
Application User Interface

🚗 About SpotWise

Parking in busy cities like Los Angeles is confusing and stressful. As a group, we often found ourselves unsure whether a spot was actually safe — unclear signs, time restrictions, and inconsistent enforcement made it easy to accidentally get a ticket. We wanted a way to answer a simple question:

“Is it safe to park here right now?”

💡 Inspiration

Our project was inspired by our own daily experiences struggling to find parking. Even when we thought we understood the rules, we still ended up second-guessing ourselves or risking a ticket.

We realized that cities already collect massive amounts of parking citation data — but that data isn’t easily usable for everyday drivers. We wanted to turn that data into something practical and accessible.

🧠 What We Built

We built SpotWise, a smart parking assistant that predicts the likelihood of receiving a parking ticket based on historical citation data.

Users can:

Enter an address or location
View a risk score (probability of getting a ticket)
See the most common violations in that area
Get actionable recommendations (e.g., peak enforcement times)

At its core, the system estimates:

$$ P(\text{ticket} \mid \text{location, time, history}) $$

This helps transform raw data into a simple, intuitive decision tool.

🛠️ How We Built It

Data: Los Angeles Parking Citation Dataset
Processing:
- Cleaned and filtered citation records
- Extracted features like violation type, time, and location
Modeling:
- Tested models such as Logistic Regression, Random Forest, and XGBoost
- Predicted probability of receiving a ticket
Frontend:
- Built a modern web interface using React
- Designed a clean dashboard for risk visualization and insights

📚 What We Learned

How to work with real-world datasets (messy, incomplete, inconsistent)
The importance of feature engineering in predictive models
How to translate technical outputs into user-friendly insights
Designing UI/UX that simplifies complex data

⚠️ Challenges We Faced

Data Quality: Missing or inconsistent location data made modeling harder
Interpretability: Turning model outputs into meaningful recommendations
Over-simplification: Avoiding results that felt “fake” (e.g., identical risk scores)
UI Design: Balancing useful information without overwhelming users

🚀 What’s Next

Integrate real-time data and APIs
Improve location accuracy (geocoding + mapping)
Add smarter recommendations (e.g., safer nearby alternatives)
Expand to other cities

🎯 Our Goal

SpotWise aims to make parking less stressful by helping users make data-driven decisions — so they can park with confidence and avoid unnecessary tickets.