Technical Assessment of Open Data Platforms for National Statistical Organisations
The term quot;open dataquot; is generally understood to be data that are made available to the public free of charge, without registration or restrictive licenses, for any purpose whatsoever (including commercial purposes), in electronic, machine-readable formats that ensure data are easy to find, download and use. National Statistics Offices (NSOs) have the potential to play a pivotal role in the implementation of open data initiatives. As producers and curators of data, the objective of making high quality data more accessible and usable is consistent with their guiding principles. NSOs indicate, in research conducted in support of this report, that one of the difficulties they encounter is that the technology they use to publish - or electronically distribute - data for public use is not compatible with open formats. They also indicate that common software packages used for open data portals do not accommodate the data formats and metadata they produce. Two key concerns related to data dissemination products are addresses: (1) Can such products designed primarily for NSOs satisfy requirements for an open data initiative?; and (2) Can such products designed primarily for open data satisfy the requirements of NSOs? Furthermore, data reuse, both by data experts and the public at large, is key to creating new opportunities and benefits from government data. The following recommendations are made to improve the overall utility of data publication platforms to NSOs and the open data community: improve technical documentation; ensure public Application Programming Interfaces (APIs) and endpoints are interoperable; presentation of metadata and Uniform Resource Identifiers (URIs) must conform to W3C standards; natural language search and metadata faceting should be standard; structural metadata and hypercube support are core NSO requirements; dashboards and visualisations are necessary for user engagement; and develop data engagement tools for improving data-quality and reuse.