FairFix

Inspiration

From personal experience and having witnessed others struggle too, car maintenence can be a difficult thing to fully understand and organize. Many people have struggled with knowing how to properly maintain their vehicle, giving it the service it needs at the right time, and most importantly, paying the right price. Dealerships, mechanics, and other shops have gotten away with charging absurd amounts or being ambigous with pricing to take advantage of uneducated customers. Some people have even reported that these mechanics will purposefully delay maintenence to make it worse in the long run, which leads to the exploitation of more money out of consumers. With this being a reality for some, we have developed an app to help combat this dillema.

What it does

FairFix combats the challenges of automotive maintenence by providing users/consumers with the best information for their car. With the input of the user's vehicle's year, make, model, and mileage, the user can choose to check what service is due next according to their specific car and mileage. They can also choose to see a specific service for their vehicle and check what a fair price is at either dealerships or indy shops. Furthermore, they can enter in their zip code to see what shops are available near them to determine and plan where they are getting their service done.

How we built it

We built a structured, data-driven maintenance and pricing platform that avoids unreliable scrapers by using curated reference datasets instead. Our system ingests OEM maintenance data (manufacturer sites and manuals), dealer and independent pricing references, standardized labor data, retailer-based parts pricing, and fully integrated NHTSA recall data. All estimates are generated from our own code and CSV-based data models, not external guesses or live scraping. Pricing is statistically modeled using mean and standard deviation to produce 95% confidence intervals, while mileage is used only to recommend services based on vehicle-specific maintenance intervals (not to adjust prices). The estimator supports dealer vs independent pricing, vehicle tiers, age-based discounts, and confidence-based ranges, and is delivered through CLI tools and a unified Flask API with a simple frontend for quotes and service recommendations.

We built the system using HTTP, APIs, JSON/CSV data, modular pipelines, REST endpoints with CORS, web technologies (HTML, CSS, JavaScript), basic statistics with confidence intervals, and YAML/env configuration, while learning to model real-world data cleanly (e.g., with Pydantic) and separate logic (estimator), data (reference CSVs), and transport (API/CLI).

Challenges we ran into

During the process, one of our group member's laptop data got fully wiped, causing our files to be lost. We spent hours fixing and recovering his laptop to continue our project. Another challenge we faced was data/output. At first, we tried getting a massive CSV file with different vehicles with different years to estimate our quotes. However, this was difficult, and caused our app to produce inaccurate quotes. We fixed this by implementing a program to first take in user input to utilize the API to scrape and find data on the user's vehicle which led to quite accurate quotes.

Accomplishments that we're proud of

Considering how overwhelming car maintenance can be, we are proud of how user friendly our app is and how easy and quick it is for a consumer to get the right information they need. We are also proud of how accurate our quotes were, and how specific they are for each vehicle. On top of that, our innovation of finding shops and dealerships around the user's area wraps everything together cleanly, providing full assistance to the user and their vehicle maintenance. Overall, we are proud of all these accomplishments and the app as a whole.

What we learned

We found that scraping OEM, dealer, and parts sites is unreliable due to layout changes, blocking, and regional differences, so we moved to reference CSVs for labor rates and parts costs, while using NHTSA’s stable API for recalls. Labor hours are fairly consistent, but labor rates and parts prices vary by vehicle tier, shop type (dealer vs independent), and location. Mileage mainly affects maintenance schedules (like different spark plug intervals by model), not pricing, so it’s used only for service recommendations. By modeling costs with mean and standard deviation and reporting 95% confidence intervals, we provide clear, defensible, data-driven estimates instead of vague price ranges, with every number coming from structured reference data or the estimator itself.

Through the process, we learned to scale ambition into something buildable: start small, choose the simplest reliable data sources, prioritize data quality over quantity, keep configuration outside the code, and focus on solving a few core problems well before expanding features.

What's next for FairFix

FairFix is already a great app with helpful features, but this is only just the beginning. There are bright futures ahead for FairFix as we already have plans of incorporating more ideas. Allowing user's to log in, and store cars in their digital garage is something that we hope to develop to make things even easier for the user. Through that, we can create checklists for each car, and even send notification reminders about services in which the user could book appointments through the app. FairFix is the gateway to removing the confusion and spontaneousness of maintaining one's car. Our goal is to make sure nobody has to feel worried or scared of maintaining their car, so that way they can pay the right price, and fix it fair.

Built With

• cursor
• google-maps-javascript-api
• html
• open-ai-api
• python
• typescript