More effective data use improves public services
Public administrations (PAs) routinely collect vast amounts of data. For privacy reasons, much of this cannot be shared. This wealth of data can however be aggregated and anonymised to generate useful statistics, such as economic and social indicators, for more informed decision-making by PAs and citizens. Linked Open Statistical Data (LOSD) – a method for modelling data – was used to good effect by the EU-funded OpenGovIntelligence (Fostering Innovation and Creativity in Europe through Public Administration Modernization towards Supplying and Exploiting Linked Open Statistical Data) project in Flanders, where a pilot converted 7 years of reporting by companies, on their emissions and waste, to the LOSD format. By linking this information to other datasets covering enterprise information, economical activities, geographical administrative data, chemical substances and others, people could use a web application, specially designed by the project, to track pollution by location. Seeing the potential, the Flemish Government’s Environment Department (Omgeving) integrated OpenGovIntelligence’s modelling approaches, components, software, guidelines and best practices into its own computer systems. In practice, this means that all the new incoming raw data produced by the department will be transformed into LOSD, with some data further transformed into alternative formats as required. The project also piloted with Trafford Council, part of the United Kingdom’s Greater Manchester Combined Authority, which was running a ‘Skills, Employment & Worklessness’ programme. Here, the team worked with the relevant stakeholders to develop a platform which could access data to support staff in their efforts to help people find employment. The platform contained: a dashboard, summarising local unemployment data; a scan application which visualised spatial distribution of unemployment; and an interactive mapping app (Signpost) that identified local unemployment resources. “By involving the people, who would ultimately use the system, at every stage of development, we helped ensure that the solution actually met their needs,” says project coordinator, Prof. Konstantinos Tarabanis.
Building the data toolkit
At the beginning of the project, the OpenGovIntelligence team realised that in order to combine datasets, and to help the web developers (who were not LOSD experts), they would have to first create middleware. This bridged front-end functionalities, used by the operators, to the back-end LOSD, which held the data. Consequently, a range of software tools for the publishing, linking and reusing of LOSD was developed, and including Table2qb and the CubiQL API. Not only did all the OpenGovIntelligence tools adopt the data modelling standard known as the ‘RDF Data Cube vocabulary’ which enables datasets of differing provenance to be combined, but the team also helped improve the standard itself, as well as establishing best practices for LOSD manipulation. Looking to the future, the team is exploring the adoption of LOSD in Artificial Intelligence (AI) scenarios. “With AI, problems and patterns could be detected faster, reducing costs and administrative burdens. Taking the Flanders example, quickly trawling through the available data, AI could more easily prioritise hotspots for inspection, improving compliance with environmental regulations and the quality of life for citizens,” says Prof. Tarabanis.
Keywords
OpenGovIntelligence, public administration, public services, citizens, data, Linked Open Statistical Data, co-creation, job seeking, pollution
