CSV on the Web
CSV is an accessible format for encoding data but it can be less than accessible for the human reading it.
CSV tables can be hard to interpret as they only have brief column headings and are shorn from context.
CSV on the Web (CSVW) is a standard for describing and clarifying the content of CSV tables.
CSVW adds a JSON “sidecar” to a CSV file that provides metadata defining a schema. This lets you provide human-readable titles and descriptions to explain your data. You can also express types, relate tables and disambiguate resources with URI templates.
CSVW is a handy technology that Open Data publishers can use to adopt FAIR principles (that data should be findable, accessible, interoperable and reusable) and support their users by treating data more like a product that a resource.
Working together with colleagues at Swirrl and the Office for National Statistics, I created csvw.org to help explain and promote the standard. I’ve also written an implementation of the standard in R that’s available on CRAN: csvwr.