Skip to main content
CUSTOMER STORY

Data-powered car-sharing gives people the freedom to drive

66%

Faster time to insights for vehicle analytics

50%

Reduction in fuel theft using geospatial analysis

20%

Faster time to market for new data products

Getgo header image
INDUSTRY: Professional,Scientific and Technical Services
PLATFORM USE CASE: Databricks SQL,Workflows
CLOUD: AWS

“Databricks has transformed how we leverage our complex data to deliver actionable insights and unlock new possibilities with AI to reinforce our mission to give everyone the freedom to drive.”

— Maximilian Jackson, Head of Data Science and Engineering, GetGo Carsharing

The Asia-Pacific region dominates the Mobility as a Service (MaaS) market, with 36.4% growth through 2025. GetGo Carsharing, Singapore’s largest car-sharing company, provides easy and immediate access to surplus fleet vehicles in over 1,500 locations. With over 300,000 drivers, 3,000 vehicles and up to 6,000 bookings daily, GetGo Carsharing generates massive amounts of telemetry, payment and image data. As data volumes continued to rise, GetGo Carsharing’s small data team struggled to manage their complex infrastructure, control and expand secure data access, and efficiently fulfill use cases such as customer 360, predictive maintenance, supply chain management and more. GetGo Carsharing migrated to Databricks Data Intelligence Platform to unify their data with analytics, helping them to apply geospatial analysis and natural language processing (NLP) for faster insights and better business solutions. GetGo Carsharing can now improve customer satisfaction and vehicle utilization with optimal placement and demand-based expansion while reducing fraudulent activity.

Struggling to drive data use and innovation in a legacy environment

Mobility services like car-sharing rely on the ingestion, analysis and application of various customer, vehicle, payment and image data sources to match available cars with customer needs. This information provides visibility for internal business units to understand user demographics and high-velocity telemetry data from vehicles. Unfortunately, GetGo Carsharing’s legacy data environment stifled their ability to manage and utilize large data sets. It also struggled to handle the influx of data from various sources including Xero, Zendesk, Esso, GetGo Carsharing’s app and real-time telemetry.

Maximilian Jackson, Head of Data Science and Engineering at GetGo Carsharing, explains, “We were just pushing data to our MySQL Community Edition database. We didn’t know who was using the data or the use cases and couldn’t govern access. I had to give out admin passwords because I couldn’t create new roles. We were downloading data, emailing it and putting it in shared drives — too much data going everywhere.” Without an enterprise-grade data solution, GetGo Carsharing could not scale efficiently, ensure data security, or implement AI and machine learning (ML).”

Additionally, because GetGo Carsharing was using a community version of their database, they needed to build and maintain their infrastructure. This required complex and manually intensive configurations, integration and management of multiple tools. In addition, maintenance tasks further slowed outcomes and increased operational costs, frustrating the data team and business alike. “It would take us one week to deliver insights to our various business units and partners,” explained Paul Bedi, Head of Partnerships at GetGo Carsharing. “This greatly lengthened time to market for critical insights that could improve our car-sharing experience — from making it easier for customers to locate and book the right vehicle to optimizing our performance and maintenance schedules.” It was time for GetGo Carsharing to harness the power of their data with a unified platform.

The lakehouse architecture lays the foundation for data and AI at scale

The GetGo Carsharing team is committed to the Databricks Data Intelligence Platform for its unifying architecture and fully managed infrastructure. As an AWS shop, they realized that relying on native services such as Redshift for data warehousing didn’t make much sense then, as they required additional effort, labor and costs to recreate what was readily available in Delta Lake, the open format storage layer at the heart of the Databricks Data Intelligence Platform. “Redshift gets really expensive really quickly, and we were thinking long term. We didn’t want to manage a bunch of AWS tools. We wanted to streamline infrastructure management to focus on AI use cases. Databricks makes that easy,” says Jackson.

Since centralizing on Databricks, GetGo Carsharing has leveraged Unity Catalog for controlled data governance and security. With visibility and authority over how data is accessed and by whom — in addition to providing simplified access on a unified, user-friendly and scalable platform — GetGo Carsharing was able to increase adoption across business data users from five to 40. The data team has also grown from five to 20 practitioners without any productivity degradation. Collaboration has only improved as data team members can now take over notebooks or collaborate with business users whenever preprocessing or complicated data transformations are required.

Using Databricks Workflows, GetGo Carsharing orchestrates data pipelines to deliver optimal performance and reliability across various data environments. Jackson says, “We have a staging branch on GitHub to test jobs in Databricks Workflows before pushing to production. With the added staging layer, we know the workflow will succeed. We’re not anticipating breaks, getting distracted by repairs or having to rerun everything in production. It’s a huge time-saver.”

With data flowing freely downstream, the engineering team can easily consume actionable insights through integrated Power BI dashboards and visualizations to monitor and improve data quality and integrity. They’re also building custom applications on top of Databricks SQL for more in-depth visualizations and leveraging Photon to accelerate data queries. Now that GetGo Carsharing has a solid foundation, it can build applications and fulfill use cases alongside business units for faster, more intelligent decision-making that benefits the customer’s car-sharing experience.

Fast, accurate and accessible analytics drive constant innovation

Since migrating to Databricks Data Intelligence Platform, GetGo Carsharing has reduced the time and effort required to build and maintain data products, resulting in a 20% boost in data team productivity. With improved data processing, GetGo Carsharing has accelerated time to insights by 66% and time to market for new insights such as fleet maintenance scheduling, fault prediction, vehicle 360 and more. Now, GetGo Carsharing delivers next-business-day insights across all use cases — a 7x improvement from their legacy environment.

They can now use vehicle telemetry data and geospatial analysis to identify regions of interest based on usage, and optimize vehicle placement and partner relationships. Classifying bookings and trips also influences the types of vehicles GetGo Carsharing adds to their fleet. Large and reliable vehicles are necessary for suburban families traveling long distances, while singles in the city want small, fuel-efficient cars. GetGo Carsharing can balance supply and demand to increase customer satisfaction, drive utilization and cost-effectively scale their fleet.

GetGo Carsharing can also minimize fuel theft that occurs through the misuse of the Esso fuel cards included in each vehicle. By looking at each user’s booking behavior and refueling patterns, GetGo Carsharing can identify fraudulent actions through disparities and take steps to address them. This has helped reduce fuel theft by 50% and increase customer satisfaction.

Moving forward, GetGo Carsharing plans to expand their use of data beyond internal analytics and introduce customer-facing features powered by machine learning and AI. Jackson concludes, “Databricks is akin to a data forge where we find various tools to fulfill use cases that solve business problems across finance, business intelligence, operations, marketing and product development. Now, we can churn out more data products, get them into production and iterate with feedback to constantly drive innovation.”