The process of retrieving data from a source system — such as a database, file, API, or website — for further processing, transformation, or analysis.
Data extraction is the first step in most data workflows. The "E" in ETL (Extract, Transform, Load) refers to this step. Extraction methods vary by source: SQL queries for databases, API calls for web services, file downloads for open data portals, and web scraping for websites that do not offer structured data exports.
The goal of extraction is to get data out of its source system in a usable form, without modifying the source. Extraction should be documented so that the provenance of the data is clear.
Example: A data analyst needs to combine property assessment data from three provincial databases. They extract data from each database using SQL queries, export the results as CSV files, and then proceed to the transformation step — standardizing formats and resolving differences before loading the combined dataset into an analysis environment.