A data source is where data comes from. Databases, APIs, log files, sensors, surveys, and public records all serve as sources. The source determines the data's structure, quality, and reliability. A well-maintained database with constraints and validation produces cleaner data than a spreadsheet filled out by hand. A sensor with calibration drift produces biased readings. A public dataset may be incomplete or outdated.

Knowing the source is essential for trusting the data. Where did it come from? Who collected it? When? How? What was the method? What was excluded? These questions determine what the data can and cannot support. Data from a survey of 100 people cannot support claims about a population of millions. Data from a system that only logs successful transactions misses the failures. Data from a vendor may be accurate for their purposes and wrong for yours. The source also affects the pipeline. A database with a stable schema is easier to integrate than a website that changes its HTML every month. An API with rate limits constrains how much data you can pull. A file that arrives manually is prone to delay and error. The source is the foundation. A weak foundation undermines everything built on top of it.

Common data sources

• Databases — structured records from applications
• APIs — programmatic access to external systems
• Log files — system and application events
• Sensors — continuous physical measurements
• Surveys — structured responses from people
• Public records — government and open datasets

Before trusting the data, understand the source. The answers are often more revealing than the data itself.