This document summarizes a presentation about processing big data for predictive analytics. It discusses how Predikto uses Apache Spark and Elasticsearch to handle large datasets and perform predictive maintenance. Spark is used for ETL, feature engineering and selection due to its ability to handle large datasets and scale horizontally. Elasticsearch is used for querying, visualizing and storing time-series data due to its fast reads/writes and ability to handle different schemas. The presentation discusses implementing these technologies, including using directed acyclic graphs in Spark and optimizing Elasticsearch for memory and storage.