End-to-End Data Science with SAS

Chapter 5: Create a Modeling Data Set

Overview

ETL

Extract

APIs

Web Scraping

Open Source Data Sets

Data Set

Get the Data

Reduce the Size of the Data

Create a Target Variable

Creating TRAIN and TEST Data Sets

Variable Selection

Transform

Load

Chapter Review

Overview

It is commonly estimated that at least 80% of a data scientist’s effort is exerted in the extract, transform, load (ETL) stage of model development. This is a critical stage of model development that is often overlooked because it is not as exciting as applying a range of awesome algorithms to your data and evaluating your model’s performance. The ETL process is critical for quality model development because of the GIGO rule: Garbage In, Garbage Out.

Nearly all data sets need to ...