Zubie’s connected car app thrives on Google BigQuery’s real-time analytics

Zubie, which provides vehicle diagnostics and feedback about driving habits via a real-time car-monitoring app, depends on Google Cloud Platform to drive its service. The company relies on Google App Engine, Google BigQuery and Google Compute Engine for the speed and performance necessary to power real-time analytics and feedback.

Providing real-time data

Most car owners have had the unsettling experience of getting the infamous “check engine light” in their car and not knowing what to do about it. Zubie has stepped into that information vacuum with a service that offers vehicle diagnostics, car-monitoring information and driver safety features. The company’s solution encompasses the combination of a hardware device called the Zubie Key which plugs into the car On-Board Diagnostic Port (OBD-II) to gather information from sensors and Engine Control Units (ECUs), and cloud-based analytics and services to provide useful information to an app on drivers’ mobile devices.

The underlying challenge is how to provide that information in near real time so it can help drivers in the moment. Zubie’s service is data heavy — the company collects and monitors data from Engine Control Units and and voltage systems, as well as data that helps it evaluate the habits of the driver, such as location, driving speed and braking rates. The company also integrates third-party data to augment the data set and provide rich actionable information to consumers, small businesses and enterprises such as insurance and automotive service centers.

Connecting, analyzing and delivering that data in real time requires quite a bit of horsepower. Says Steven Zeng, Zubie’s director of solution architecture & IT: “It’s a complex ecosystem built around speed and instant information, and the pace of change is only going to increase. Zubie can’t afford to use a platform that is less than rock solid in terms of scalability and performance.”

High-performing and affordable technology

The company turned to Google Cloud Platform to support its performance and analytical needs.

Zubie uses several components of GCP. Google App Engine serves as the front end and database, retrieving data from the Zubie Key and mobile app and storing it in App Engine’s datastore. From there, the company pulls data into Google BigQuery for real-time analytics. Zubie also uses Google Compute Engine to run software and tools that cannot be run on App Engine in a scalable environment, as well as Google Cloud Storage to exchange data between different systems and archive data.

Cloud Platform fits the company’s budget without forcing a performance compromise. For example, the company’s BigQuery implementation required no upfront technology investment, and took only about two and a half days to set up. “I’ve worked on analytics projects at other companies where we needed a minimum of $250,000 just for software and hardware,” Zeng says.

Boosting developer productivity

Zubie’s business value depends on its ability to convey accurate answers to its enterprise customers and general consumers, and BigQuery is instrumental to the process. “We sample some of the data elements at a per-second rate; BigQuery can analyze that data within a few seconds without any database tuning,” says Zeng.

BigQuery also helps the company understand and adjust to data inconsistencies that arise from a lack of industry standards around vehicle information. For example, BigQuery parses diagnostic data that differ by vehicle maker and model. Says Zeng: “By understanding the differences in how each vehicle behaves from a data standpoint, our development team can adjust for these quirks and give accurate scores in real time.”

GCP’s flexibility and ease of use also improves developer productivity, allowing the company to react more quickly. For example, Zubie’s early data models changed weekly as developers honed in on customer preferences and strategic directions. “We were able to act nimbly because with App Engine we didn’t have to worry about infrastructure. And with NoSQL, modifying data models was a snap — data model changes made in the code will be automatically persisted in the data store,” says Zeng. “This made changing data schema and deployment very easy.”

App Engine features like the user interface and command lines for major features also accelerate productivity for developers at all levels. “Even developers without much experience can generally start coding in the first month or so,” says Zeng.

GCP’s cutting-edge infrastructure lets the company devote more resources to building core app features rather than to IT maintenance tasks. “Cloud Platform has been invaluable in helping our engineers focus on the core features of our product by removing common IT maintenance tasks, and it does so at a good price,” says Zeng.

Integrating important aspects like scalability and high availability into the platform also improves IT efficiency. Take, for example, the scaling feature on App Engine. “Having something that I know will auto scale puts me in a very comfortable zone—we don’t have to spend much time researching how to respond in the event of a traffic spike,” Zeng says.

By maximizing the productivity of Zubie’s most valuable resource, workforce talent, Google Cloud Platform helps Zubie more quickly expand its product and market into mainstream acceptance.