Inspiration: Tix Why I Built It
Last winter, I got a parking ticket that honestly made no sense. I was parked on what I thought was a completely legal street. No obvious signs. No blocked driveway. No expired meter. Just a bright orange envelope on my windshield. $115. I remember standing there thinking Is this even valid? Did I miss something? Do I actually have to pay this? I ended up paying it not because I believed it was right, but because I did not know how to challenge it. Later I learned that I was far from alone. Major U.S. cities issue hundreds of thousands to millions of parking tickets every year. In dense cities like New York, enforcement is constant and the rules are complex. For most people, disputing a ticket means navigating forms, deadlines, hearings, and potentially weeks or months of waiting as well as wasted money. That uncertainty is what Tix is built to reduce.
The Problem
Parking enforcement is data driven, but the average driver has no access to that data. Cities like New York publish detailed parking violation records publicly. Millions of ticket records including violation codes, timestamps, locations, issuing precincts, and outcomes are available through open data portals. But that information is buried in spreadsheets and databases that most people will never look at. At the same time, the dispute process can be intimidating. In many cities, you must file a challenge within a short window or lose your right to contest. If you pursue it further, it can involve multiple review stages and extended timelines. So people default to paying.
What Tix Does:
Tix is an AI powered parking ticket analyzer. You upload your ticket information and it checks for: Missing or incorrect required fields Formatting inconsistencies in plates, states, and timestamps Logical contradictions Violation code mismatches Structural weaknesses in how the ticket was issued
Then it outputs: A Pay Likelihood Score A Fight Likelihood Score A recommended action Pay now Fight it Ask for reduction Request a hearing
It does not promise that you will win. It gives you clarity before you decide.
Why I Trained It on New York Data: It is public
New York publishes detailed ticket level data through its Open Data portal. That means real enforcement records, real violation codes, real patterns at scale. Instead of inventing hypothetical examples, I was able to analyze millions of actual tickets. That matters for three reasons. First is scale: New York issues millions of parking tickets each year. That provides enough data to identify patterns, common error structures, and enforcement trends. Second is transparency: The data is publicly accessible. Anyone can see the same dataset. We are not using private records or hidden systems. We are building on open civic data. Third is complexity: New York parking rules are notoriously dense. Alternate side parking, street cleaning schedules, loading zones, commercial vehicle rules, dozens of violation codes. If a system can reason through New York’s structure, it creates a strong foundation for expanding to other cities. Tix is designed to be modular. As more cities publish structured ticket data, the same validation logic can be adapted and retrained.
How I Built It
I built Tix as a web application using Next.js and codex Under the surface, the core is structured validation logic combined with a scoring system. Each ticket is evaluated through rule based checks. Those checks contribute weighted signals toward a fight likelihood score. The more structural inconsistencies or mismatches detected, the higher the fight score becomes.
What I Learned
Most people do not fight parking tickets because the process feels intimidating (Spoke to family, friends, and interviewed people in the village). Small technical details matter more than people realize. Also, real world data is messy. Fields are inconsistent. Codes change. Edge cases are everywhere. It is hard to find good training data.
Challenges:
Translating legal style rules into clean, explainable logic was harder than expected. Balancing statistical reasoning with transparency was also hard and took a lot of analysis. Also, debugging was a pain.
Why This Matters
Millions of parking tickets are issued every year. Most are paid without question. Some deserve to be paid. Some do not. Tix does not promise to eliminate tickets. It gives people information before they make a financial decision. Sometimes the right answer is pay. Sometimes it is fight. But it should be a decision made with clarity, not confusion. And that is what Tix is built for.
What's next for Tix
Right now, Tix is trained primarily on New York City parking violation data because it is publicly available, structured, and large enough to build meaningful validation logic. The next step is expanding the training data to include more cities. Many major cities publish open parking enforcement datasets. By incorporating data from places like Los Angeles, Chicago, San Francisco, and others, Tix can: Improve pattern recognition across jurisdictions Adjust validation logic to local violation codes Learn city specific enforcement behaviors Increase accuracy in fight likelihood scoring
Log in or sign up for Devpost to join the conversation.