Reducing car maintenance times with near real-time analytics

About:

Kovi offers a holistic approach to car leasing for gig workers with services that cover financing, insurance, and maintenance. Founded in 2018, Kovi’s mission, as articulated by CEO Adhemar Neto, is to “make car ownership affordable, efficient, and accessible to all, by building a scalable platform that creates value for all.” To ensure operational efficiency and deliver optimal customer experiences, Kovi relies heavily on near real-time analytics. For example, Kovi has in-car technology to analyze, among other things, driver’s driving behavior, as well as in-house maintenance operations that analyzes how to reduce idle time. The company’s platform, built on AWS infrastructure, provides the scalability, flexibility, and automation required for its operation, which handles over 2 billion data points daily.

Challenge: 

Kovi’s previous reliance on Amazon Athena for analytics posed several challenges, particularly in the transformation time in their medallion architecture, which took an hour. This delay not only impeded the company’s ability to deliver timely car maintenance to its customers, but also increased costs and affected customer satisfaction. Johni Michels, Data Engineering Team Lead at Kovi, expressed frustration with the situation, stating, “The delay in analytics with Athena was crippling our operations. It took us an hour to transform data from Silver to Gold layer, making it impossible to deliver car maintenance within our 3-hour goal.”

Solution: 

To address these challenges, Kovi evaluated several options, and ultimately decided to employ Starburst Galaxy and Iceberg to scale its data lakehouse. Kovi chose Starburst Galaxy for its superior price/performance ratio, which resulted in faster insights at a competitive cost. Johni highlighted the decision, stating, “We chose Starburst Galaxy because it offered the best price/performance, allowing us to accelerate our analytics process while optimizing costs.” The decision to select Starburst over alternative options was also influenced by its Warp Speed feature for accelerating ad-hoc queries and fault-tolerant executionmode for faster ETL jobs. Additionally, Iceberg was chosen for its efficient data management capabilities, providing scalability and reliability for Kovi’s growing dataset.

Results: 

Since implementing Starburst Galaxy and Iceberg, the following improvements have been documented:

• Reduced data transformation delays from 60 minutes to just 10 minutes, an improvement of over 80%.
• 75% reduction in AWS S3 GET costs and 50% reduction in Fivetran costs, a significant cost savings.
• 85% faster ad-hoc queries with Warp Speed and 55% faster ETL jobs with fault-tolerant execution mode, an increase in operational efficiency.

With a more performant lakehouse, Kovi transformed its analytics processes to enable near-real-time insights, which resulted in significant improvements in operational efficiency, including an estimated 10-20% reduction in car maintenance downtime. Additionally, Kovi can now track geographical data more effectively, which reduces the number of safety-related incidents.

Overall, Starburst Galaxy has empowered Kovi’s business teams to ask their own questions, formulate new ideas, and drive innovation, fostering a culture of ownership and collaboration across the organization. “The primary achievement from Starburst and Iceberg is that our business teams are more autonomous,” shares Iuri Zambotto, Data Engineer at Kovi. “This creates a shift in the organization where we align data analytics with our business objectives from the start, and as a result we’re able to innovate faster to provide better value to our customers.”