A JSON format to describe civic data resources & microservices

microcivic.json

{
	"name": "phl-property-tax-balances",
	"keywords": ["tax", "property", "delinquent", "deadbeat"],
	"license": "CC0",
	"sources": [{
		"name": "City of Philadelphia",
		"web": "http://opendataphilly.org",
		"email": "[email protected]"
	}],
	"resources": [{
		"name": "Property tax balance API",
		"url": "https://data.phila.gov/resource/y5ti-svsu.json",
		"format": "json",
		"encoding": "UTF-8",
		"authentication": false,
		"ratelimit": true,
		"documentation": "https://www.opendataphilly.org/dataset/property-tax-balances/resource/927260a9-1e73-41c7-8553-9061d0334f47"
	}, {
		"name": "Property tax balance download",
		"path": "/realestate-tax-balance.csv",
		"format": "csv",
		"bytes": "",
		"hash": "sha1:8843d7f92416211de9ebb963ff4ce28125932878"
	}],
	"schema": {
		"fields": [{
				"name": "interest",
				"type": "number"
			},

			{
				"name": "lien_number",
				"type": "string"
			},

			{
				"name": "location",
				"type": "string"
			},

			{
				"name": "other",
				"type": "number"
			},

			{
				"name": "owner",
				"type": "string"
			},

			{
				"name": "parcel_number",
				"type": "string"
			},

			{
				"name": "penalty",
				"type": "number"
			},

			{
				"name": "principal",
				"type": "number"
			},

			{
				"name": "tax_period",
				"type": "number"
			},

			{
				"name": "total",
				"type": "number"
			}
		]
	}
}

My feeling is that this format should live a level above the data packages standard, with the latter one of the standards that could be referred to. For example, something like schema would not appear here. Either this would refer to a standard that itself specifies where the scheme is to be found (data packages) or it just specifies a package file with that information.

Not sure exactly how to represent, but other things of interest would be:

• maturity: Some measure indicating adherence to quality standards
• validation: pointer to testing/validation resources
• use: pointer to or even just included usage resources (e.g., wrappers to get, display, test, etc.)