PipelineDB

Robyn is an open source, experimental Marketing Mix Modeling (MMM) package developed by Meta’s Marketing Science team. It’s designed to help advertisers and analysts build rigorous, data-driven models that quantify how different marketing channels contribute to business outcomes (like sales, conversions, or other KPIs) in a privacy-safe, aggregated way. Rather than relying on user-level tracking, Robyn analyzes historical time-series data, combining marketing spend or reach data (ads, promotions, organic efforts, etc.) with outcome metrics, to estimate incremental impact, saturation effects, and carry-over (adstock) dynamics. Under the hood, Robyn blends classical statistical methods with modern machine learning and optimization; it uses ridge regression (to regularize against multicollinearity in many-channel models), time-series decomposition to isolate trend and seasonality, and a multi-objective evolutionary algorithm.

Anomaly detection in time series data is essential for the day-to-day operation of many companies. With Timeseries Insights API Preview, you can gather insights in real-time from your time-series datasets. Get everything you need to understand your API query results, such as anomaly events, forecasted range of values, and slices of events that were examined. Stream data in real-time, making it possible to detect anomalies while they are happening. Rely on Google Cloud's end-to-end infrastructure and defense-in-depth approach to security that's been innovated for over 15 years through consumer apps like Gmail and Search. At its core, Timeseries Insights API is fully integrated with other Google Cloud Storage services, providing you with a consistent method of access across storage products. Detect trends and anomalies with multiple event dimensions. Handle datasets consisting of tens of billions of events. Run thousands of queries per second.