Improving CarVertical's Vehicle History Reporting with Data Labeling

CarVertical, a leader in vehicle history reports, helps individuals and businesses verify car authenticity, accident history, and service records. To further improve the accuracy of their reports, CarVertical partnered with Sapien to streamline data labeling processes for catalog lookups, VIN tagging, and vehicle image positioning. This collaboration focused on refining how vehicles were matched to user queries, how VINs were extracted, and how images were categorized.

The Challenges

CarVertical faced several challenges in improving their data accuracy. They needed a system that could more efficiently match user queries to the correct vehicle makes and models, extract and label VINs accurately, and position vehicle images consistently in their catalog. The goal was to increase the speed and precision of their processes, ensuring customers received reliable vehicle history information with minimal effort.

Challenge: Prompt Matching

To help CarVertical match vehicle queries to the right car models, Sapien developed a product-matching tool that could accurately filter out irrelevant search results. This system used a catalog lookup assistant to match user-generated prompts to specific vehicle brands and years. The tool was designed with a focus on automation, simplifying the labeling process for labelers. We also incorporated a labeling platform that pre-labeled data, which greatly reduced the manual workload for labelers.

Outcome

Sapien delivered a highly accurate prompt matching tool that not only minimized search errors but also streamlined the entire search process. This helped CarVertical build a more efficient and user-friendly catalog lookup system, enhancing their overall vehicle history reporting.

Challenge: VIN Labeling

Accurate Vehicle Identification Number (VIN) extraction is an important element for CarVertical’s reporting system. To improve this, Sapien trained Optical Character Recognition (OCR) tools to identify and extract VINs automatically from thousands of VIN tags. This process included designing an automatic detection feature that simplified the task for labelers. The result was faster VIN matching and more accurate data extraction.

Sapien also introduced a Bounty Program that rewarded our global decentralized team of labelers with points for completing VIN tags and other tasks. This initiative motivated the labeling community to engage more actively, increasing the speed and efficiency of the process.

Outcome

CarVertical saw significant improvements in their OCR capabilities. The automated VIN detection feature helped labelers work faster and more accurately, leading to quicker data lookups and better reporting for CarVertical’s customers.

Challenge: Vehicle Image Positioning

CarVertical needed to standardize the way vehicles appeared in their catalog, with consistent angles for each car (front, side, rear, etc.). Sapien labeled thousands of vehicle images, implementing a quality control process to ensure each image was correctly categorized. This made the catalog more visually coherent and laid the foundation for future advancements in automated vehicle recognition.

Outcome

CarVertical’s vehicle catalog became easier to navigate and more visually appealing. The consistent image positioning improved the user experience, and the accurate labeling of vehicle angles provided valuable data for potential future applications in automated vehicle systems.

Gamification and Its Impact on the VIN Labeling Project

One of the standout features of our collaboration with CarVertical was the introduction of gamification into the data labeling process. Understanding the challenges and monotony that often accompany data labeling tasks, Sapien’s gamified platform aims to increase  engagement and productivity among our global decentralized community of labelers.

Our Bounty Program was at the core of this gamified experience. Labelers earned points for completing specific tasks, such as VIN tagging and cataloging vehicle images. These points were were part of a broader engagement system that motivated labelers to contribute consistently and efficiently. The points earned can be redeemed for crypto rewards, creating a sense of competition and achievement that kept labelers highly motivated.

Thanks to this system, overall labeling speed and accuracy improved significantly, allowing CarVertical to enhance their vehicle history reporting and deliver precise, reliable information to their customers.

Conclusion

Sapien’s collaboration with CarVertical significantly improved their vehicle history reporting by introducing an efficient labeling platform, automating VIN extraction, and enhancing vehicle image positioning. Our decentralized, precise data labeling approach was absolutely instrumental in fine-tuning the AI tools used by CarVertical. The introduction of the Bounty Program also streamlined the labeling process by incentivizing labelers. Through this partnership, CarVertical improved their ability to deliver accurate, reliable vehicle information to customers, elevating their service to completely new levels.