SemStats 2016 Program

The concept of “evidence-based policy making” (EBPM) has been gaining currency over the past decade as online open government databases, together with sensors and mobile devices that enable systematic monitoring, make it increasingly viable. Although EBPM appears logical, one of the reasons it has not been widely adopted is because of the difficulty in acquiring, integrating and analysing relevant benchmark and evidential data. Successful EBPM depends on the capture, integration, processing and analysis of a wide range of data and knowledge (including policies, programs, targets, indicators) that are heterogeneous, unstructured, qualitative and quantitative, inter-dependent and dynamic. Moreover, different agencies and regions are defining and employing many similar but different sets of indicators (e.g., Global City Indicators, UN Global Urban Indicators, World Bank Indicators, Millennium Development Goal Indicators) to measure a wide range of attributes including: economic development, urban development, sustainability, Quality of Life, environmental health, educational standards etc.

This paper will describe the critical components required to underpin an evidence-based policy-making ICT framework and support interoperability between indicators across programs. It will also describe the ontologies and services that are being implemented, applied and evaluated within a number of projects in Australia to support evidence-based policy making within urban environments, including:

• Ontological representations of indicators that include: calculation methods; units; statistical type; and relationships to other indicators,
• A collaborative ontology editing interface that allows users to search, browse, retrieve and tailor existing indicators to define new indicators;
• A high level EBPM ontology that links policies, to programs to indicators and raw data – and bridges the gap between the agencies who collect the monitoring data and the policy-makers who consume it;
• APIs and SDMX-based interfaces that enable dynamic retrieval of distributed data sets that underpin the indicators;
• Web interfaces that enable users to interactively select indicators and contexts of interest, and dynamically generate, display and overlay indicator values, spatially and graphically;
• Spatio-statistical analysis services to identify and record correlations and interdependencies between different indicators or parameters.