A data source is a system, repository, or collection that supplies data for an analysis or visualization. It might be an internal warehouse table, a spreadsheet, a third-party service, or a public dataset. Analysts often combine sources to answer a business question that no single dataset can address. section 10.0 section 11.2.1
Choose a source by asking what the decision requires and how the data will be maintained:
- Does it contain the needed information at a useful level of detail? Public economic or demographic data can add context that an organization does not collect itself, but availability varies by location and topic. section 10.3 section 10.3.2
- How quickly will it become outdated, and how often must it refresh? A regularly used, changing source may justify a warehouse pipeline; a relatively static or short-lived source may be uploaded manually. section 10.1.3 section 11.2.1
- How can it be retrieved? Options include an existing connector, a custom API pipeline, or a manual download and upload. Consider the effort alongside how widely the data will be used. section 10.1.3 section 11.2.1
Record where each source came from and how its data was retrieved, shaped, and processed. Keep final queries, useful exploratory queries, third-party links, and retrieval code. This history lets others rerun the analysis, investigate errors, and understand why some datasets were excluded. For a stakeholder presentation, a short source summary can explain both the datasets used and those left out. section 2.3.2 section 2.3.3
Choosing data for a location decision
A sustainable-clothing retailer considering a new city could explore census and survey data for age, income, and concern about environmental sustainability. Before recommending a location, check which source actually supports each comparison and decide whether future monitoring calls for an API pipeline, a manual CSV download, or periodic checks of a data explorer. section 10.3.4
