Retail data science company

Improving data science

The parent company built a vast data lake from varied sources including customer loyalty information, point of sale transactions, supply-chain logistics and store performance metrics. The key to leveraging this data to improve business performance was to provide the company’s data scientists with quick and easy access. However, the current platform included many manual approval and provisioning processes, which slowed access, and a user interface that didn’t allow for full visibility in other lines of inquiry and analysis. Making these processes faster, better integrated and more efficient was crucial for the parent company to hone its competitive advantage over new arrivals.

Connecting the dots, faster

Catalyte engineers hit the ground running to help the retail data science company reduce delay and increase performance of its data science platform, and implement best practices for development and architecture. Aggregating multiple background services, Catalyte engineers built a custom, intuitive user interface for the client’s data scientists. This new UI allowed data to be more easily searched and analyzed. It also created searchable metadata that advanced collaboration among the data scientists. Code and analysis outputs could now be shared with, or searched by, other internal teams. Built into the new UI and aggregation systems was a level of primary user identification/authentication. This eliminated the need for manual authentication, which had slowed down access to data, and increased the security surrounding the data lake.

Making the data lake easier to navigate

Catalyte’s work allowed the data science platform to deliver critical information and insights faster, resulting in better informed business decisions. With a new on-demand platform, data scientists could set up infrastructure in a matter of hours, rather than months. The platform also provided a uniform, single source of data truth with granular permissions, simultaneously increasing collaboration and security. Now with the ability to better find and reuse queries and data models, the company can produce data science at the speed of modern business comparable with other digital-first competitors.