Case Study: Facebook achieves time-series forecasting and in-chart analytics with Preset Superset

Facebook enables time-series forecasting in 0.38 release with Preset

The customer, Facebook, sought to democratize the use of advanced time-series analysis and forecasting for its users, making powerful tools accessible beyond data scientists with specialized backgrounds. The challenge was to integrate a highly approachable forecasting package, Facebook's Prophet, into a user-friendly data visualization environment, allowing analysts to make predictions with minimal tuning. Vendor Preset addressed this by leveraging its deep integration with the Apache open-source ecosystem.

Preset implemented a solution using Apache Superset, which was enhanced with experimental support for the Prophet library. This integration allowed users to perform in-chart analytics and time-series forecasting directly within the Superset visualization interface. By connecting Superset to the time-series database Apache Druid, Preset created a powerful stack for fast data ingestion and retrieval. While the case study focuses on the technical setup, the result is a democratized analytical environment where users can easily apply forecasting models, paving the way for more accessible data insights.

Facebook