How We Built Nucleus: Community Brands' Analytics Platform

Community Brands is a leading technology company for associations, member-based organizations, nonprofits and schools (K-12). We rely on the Nucleus Analytics platform to provide customers an interactive, unified view of business’ most relevant and time-sensitive data.

In this talk, we describe how the Nucleus engineering team built a real time, user-facing analytics app with the following requirements:

• Data latency of under two seconds for real-time data
• Sub-second query latency for user-facing analytics
• A fully-managed cloud architecture for operational simplicity
• A fully mutable database capable of handling updates from transactional systems

Using Confluent’s cloud-native service for Apache Kafka, Apache Flink, Rockset, and transactional databases in our architecture, we can now quickly and easily serve real-time analytics to our customers via change data capture (CDC) streams. Since switching to Rockset, our operational overhead has decreased by an order of magnitude. In this talk, we will share how we designed and architected our real-time analytics platform, how we use real-time CDC, and the principles used to build new data products with greater speed and agility.