Access to public open space is an important determinant of levels of physical activity. Recreational use of parks and public open space promotes better cardiovascular and mental health. Increased levels of physical activity help prevent non-communicable diseases such as cardiovascular disease and diabetes. It is therefore important that planners consider public open space in the design of our cities.
Most state governments in Australia have policies aimed at providing city residents access to parks and public open space. In Melbourne, the Victorian Planning Provisions state that there should be “local parks within 400 m safe walking distance of at least 95% of all dwellings”. Local parks are defined as either including active open space (e.g. an oval) or otherwise being generally 1 hectare (Ha) in area.
Open data and free and open source software for geospatial (FOSS4G) jointly provide us with the opportunity to measure and map which dwellings in Melbourne achieve this policy. It is therefore also possible to assess whether the 95% target is being met in suburbs, local government areas, and Melbourne overall.
There are multiple open data sets that could be used to conduct the analysis. The Geocoded National Address File (G-NAF), available at https://data.gov.au/dataset/geocoded-national-address-file-g-naf, gives us all address points in Melbourne. The ABS provide city, local government area and suburb boundaries, along with small area (Mesh Block) dwelling counts. Through Victoria’s open data directory (https://www.data.vic.gov.au/), Vicmap provide property centroids and road networks through and the Victorian Environment Assessment Council (VEAC) provide public open space data. Road network and public open space data is also available through Open Street Map.
Using a range of free and open source software for geospatial (FOSS4G), our team have calculated public open space access for each and every dwelling in Melbourne for different combinations of these data sets.
So does our choice of open data sets matter? When we measure policy implementation, do we get a different result when we use crowdsourced data versus official government data? In this presentation, we will show that the answer to both these questions is ‘yes’. We will show which areas of the city differ in their results, and by how much, depending on the data sources used. We will examine what is causing these different results and provide an assessment of which data sets provide the most accurate results in different circumstances. We will also make recommendations for the improvement of official government data sources and show that our findings have significance in other Australian states and territories.