Saturday, 20th April 2019
Logo

Powering a social fitness network

Strava, the GPS tracking app and social network for athletes, uses Snowflake to power its business. Switching to Snowflake has significantly reduced the data engineering resources required to manage the company’s data, resulting in more engineers focusing on developing additional services for Strava’s athletes and more focus on building its global community.

Since its founding in 2009, 35 million cyclists, runners, hikers and other athletesin 195 countries have joined the community of San Francisco-based Strava (Swedish for “strive”). The company’s free mobile apps and website help members discover and plan workouts, record and share activities, and analyse and compare performance. Strava’s ability to connect athletes around the world makes fitness a more social experience, providing a source of motivation and inspiration even when training alone.

“Fundamentally, we now have all of our data in one place so that we can get a true global view of how enthusiasts and athletes are using Strava,” Sr. Director of Analytics and Data Science, Cathy Tanimura said. “Snowflake enables us to draw new connections between communities so we understand how and why they enjoy using Strava, and that allows us to offer them new features that would not have been possible or practical before.”

Strava started using Snowflake in early 2018 after running into concurrency issues, among other problems, with its previous data warehouse solution. When Strava started tracking how long queries would take, the company noticed hundreds of queries per week took more than 20 minutes. Finding that unacceptable, they did an assessment of data warehouse solutions and chose Snowflake. Important factors in the decision were Snowflake’s ease of use and administration.

The key Snowflake benefits include:

Decision-making – Snowflake provides more visibility than Strava’s previous data warehouse solution into which features customers use and how they want to use them. This helps Strava determine which new features to release “into the wild,” as Tanimura describes it. Analysis performed by Snowflake helps them make better and faster decisions about what to offer its users. Snowflake also makes one-off, aggregate computations easier, which supports initiatives such as the end-of-year video for each of the athletes in their community.

Managing data – With a 120 TB data warehouse, 13 trillion GPS data points, 15 million uploads/week and 1.5 billion analytics points ingested daily, Strava needed data warehousing capabilities that only Snowflake provides. Some of the critical features include consolidating all data in one place, rapid analytics, and easily loading, integrating and analysing all data formats – structured and semi-structured. Strava’s athletes send data from all around the world, every second. The company also pulls from the data feeds of third-party software services such as Google and Facebook to provide insight into what’s going on with all aspects of its business.

Community building – By using Snowflake to create a single source of truth with its data, Strava can garner deeper insights about what’s happening on its platform. Snowflake enables Strava to draw new connections between communities so the company understands how and why the community enjoys using the apps.

“With Snowflake we started seeing value immediately,” Tanimura said. “In fact, a query which had been giving us fits for months with our previous solution was resolved in a matter of hours by Snowflake. There is no fuss and no drama with Snowflake, which is exactly what we were looking for.”

“Strava came to us with data warehousing needs that went well beyond what legacy products and cobbled-together solutions could support,” Snowflake VP of Product, Christian Kleinerman said. “Its sophisticated use cases and massive data sets needed a unique and comprehensive solution to allow them to scale while simultaneously exceeding the expectations of their millions of users. Strava is a great example of why we started Snowflake, and we’re looking forward to growing alongside them.”

Dimension Data and Cisco have reached a new co-innovation agreement, to develop a deeper collaborati...
New smart city platform ensures city control, policy transparency and citizen trust.
New advances make it easy to deploy native Spark and Drill applications in Kubernetes.
Medco Power Indonesia and NEC Corporation will begin proving trials as of April for a system to dete...
BlackBerry Cylance has published the findings of its global artificial intelligence (AI) security su...
Lack of flexibility and scalability top legacy database issues.
Databricks has opened its new EMEA headquarters in the former House of Fraser offices on Baker Stree...
DataFest19 to showcase Scotland’s leading role in data science and AI at over 60 events from 11-22 M...