Founded nearly 70 years ago, Werner Enterprises is a North American transportation and logistics leader that operates a fleet of almost 8,300 trucks and 30,000 trailers out of 16 terminals across the United States. The company generates a massive amount of data on the constantly changing, real-time location of each of its assets. Collecting and analyzing this geospatial data is vital for smart decision-making. How often are the trucks stopping? How can they reach their destination in the shortest amount of time? Where are the best places to locate new terminals based on where customers or professional drivers are located? With the help of Snowflake and CARTO, Werner gained geospatial data capabilities that empowered data-driven decision-making and led to substantial savings in time and costs.

Performance and scale

Prior to the Snowflake implementation, Werner’s legacy data warehouse was challenged to process the amount of incoming data and scale as needed. Each truck sends a location ping to the system every five minutes, which results in more than 2.5 million pings a day. Because of the limitations of the company’s on-premises warehouse, the data was inaccessible and unusable. Running queries was time-intensive, and more complicated analyses weren’t possible. “The legacy system was very limited in terms of how we could use the GPS data from our trucks,” says David Cavanaugh, a data scientist at Werner.  

With Snowflake, Werner gained better data performance and scalability. For example, the company must report vehicle mileage for various purposes, including state tax requirements. Previously, the company could store only one month’s worth of location data, so managers could not run mileage reporting for more than one month at a time. The report remained static and historical data could not be queried. Now, thanks to the Snowflake Data Cloud, Cavanaugh and his team can quickly and easily query multiple years worth of historical data. “The performance was much better and we could actually run the queries we wanted to run,” says Cavanaugh. “I use the ping data set almost every day to answer queries from different departments.” 

Due to the previous warehouse’s lack of scalability, Cavanaugh and his team had to set rules that reduced the amount of data in the system, limiting insights. Now, with a massive amount of historical data available for querying, the team can see granular data, such as how long trucks are parked at a terminal at any given time. Increased visibility into the geospatial data helps the business better understand terminal utilization and traffic flow.

Visualization functions

Cavanaugh leads his team in integrating the new features Snowflake has rolled out since Werner’s 2020 implementation. His team uses Snowflake’s GEOGRAPHY data type, which models Earth as though it were a perfect sphere, and has also tried the new GEOMETRY data type, which represents features in a planar coordinate system. “Any time you have to pull data out of a database, you’re losing performance,” says Cavanaugh. “It’s very helpful to calculate the intersection of two polygons or areas without having to move the data.” Now, with Snowflake, Werner is equipped to use any type of vector geospatial data.

Werner also uses Snowflake partner CARTO to access geospatial visualization, analysis, and app development functionalities inside the Snowflake Data Cloud. With CARTO’s native connectors, Cavanaugh’s team can write a query and connect it to easily and quickly create visualizations—no copying and pasting, reformatting, downloading, or uploading needed. CARTO visualizations also help the business quickly answer questions on everything from trucking and transportation logistics to real estate purchases or legal and regulatory issues. For example, within a few hours, Cavanaugh easily mapped out the disruption that would be caused by an interstate road closure and sent the operations team a link to the map. This enabled operations to quickly understand the impact of the closure, select the best truck routes and detours, and proactively alter delivery schedules.

A game changer

Cavanaugh’s team has also used Snowflake Marketplace to access geospatial data more easily in areas such as property boundaries, census regions and zip codes. Previously, the team had to spend time manually collecting and updating that data. 

Finally, Cavanaugh and team were able to use Snowpark to create a coordinate reference conversion rate with Python. According to Cavanaugh, “Snowpark is a game changer because it gives us the flexibility to write and use our own functions.” 

Ready to learn more?

To learn more about Snowflake’s geospatial capabilities, visit our reference documentation.