Workflow Automation Templates
A library of ready-to-use workflow templates to accelerate your data journey
Moving Average FE Spark
Compute moving averages with Spark
Overview
This workflow processes transactional data using Spark to compute global moving average features. It generates user-level trend insights such as average transaction counts, purchase gaps, sales, and unique user activity over defined time windows, supporting deeper temporal analysis.
Details
The workflow starts by loading the Moving Average Features dataset containing transaction-level data. Using the Moving Average Features node, it calculates rolling averages for amounts, transaction counts, quantities, and user activity across multiple time frames.
The resulting dataset includes both the original input and newly generated moving average columns that capture patterns like frequency, recency, and value trends. The output is displayed using Print N Rows for review.
This enables scalable trend analysis and time-based feature engineering for predictive modeling and user behavior analytics.