August 2021 Summaries

Cube.js v0.28 has been released with numerous improvements, including reliability enhancements to database drivers and a raft of updates to the Cube Store. To aid developers in selecting suitable data visualization tools, an open-source list called awesome.cube.dev has been created. Additionally, a video workshop on mastering pre-aggregations is available, while Cube Cloud, a fully-managed hosting service for Cube.js apps, is nearing its general availability with design partners needed to test the pre-release product. The company has also secured $15.5M in Series A funding and highlighted community accomplishments. Cube invites developers to showcase their projects and participate in its community, with resources available on Slack and Twitter.

BigQuery is a powerful data warehouse known for handling large datasets, but its distributed architecture results in slower response times, making it less ideal for applications requiring rapid data retrieval, particularly for dynamic dashboards. To address this, an acceleration layer using Cube.js is proposed to enhance performance by storing small, aggregated tables, known as pre-aggregations, in Cube Store. This setup allows sub-second response times by serving queries from pre-aggregations instead of directly accessing BigQuery. A step-by-step guide explains how to configure Cube.js to connect with BigQuery, create data schemas, and set up pre-aggregations, enabling improved performance for dynamic front-end applications. A demo using the public Hacker News dataset showcases how pre-aggregations significantly boost performance compared to querying raw data, with source code available on GitHub for further exploration.

FeedMe, a Malaysian-based company that provides Point of Sale (POS) systems for the food and beverage industry across South East Asia, was struggling to maintain and scale their previous reporting system, which relied on SQL queries and JavaScript to process data from MongoDB, CouchDB, and BigQuery. To address this issue, they took a phased approach to deploy Cube.js, a cloud-based data platform that allows users to create custom reports. The FeedMe team found Cube.js to be relatively straightforward to use and implemented its Dynamic Schema Creation feature first, followed by more advanced features such as pre-aggregations and Cube Store. With Cube.js, FeedMe was able to simplify their data processing, reduce their BigQuery costs by about ⅔, and improve the performance of their sales reports, which now take less than 10 seconds to query compared to several minutes before. The company has been using Cube.js in production since May 2021 and values its collaborative community and documentation as essential tools for onboarding and feedback.