Intelligent Intervention for Data Analyst Workflow

Choosing how to analyze a specific piece of data can be more of an art than a science. Often, analysts draw on years of experience to determine what tests to run and what models to build for different types of data. In situations where analysts are presented with data that they are not familiar with, it would drastically speed up the time it takes to gather actionable insights if analysts had access to a system which could provide intelligent recommendations about types of analyses to perform on this data. In this project, I am proposing to build machine learning models that can automatically generate intelligent interventions describing next steps to take when performing data analysis on complex datasets. These interventions would be based on the data being analyzed as well as the results of previous analyses. These models will be built based on examples of analyst decision making and presented in automatically generated natural language.