Predictor Factory is a tool for data scientists, who want to use data stored in relational databases in their models. After defining the target(s), subject identifier (commonly customerID) and time of prediction, Predictor Factory crawls the database and produces a single table, which contains target(s), subject identifier, time of prediction and predictors. This produced table can be processed as usual. You can perform variable transformation and feature selection or directly jump to classification/regression in your favorite tool. All transformations are done in SQL for straightforward deployment. Supported databases: Oracle, Microsoft SQL Server, SAS, PostgreSQL and MySQL.