The Sustainable Development Goals is an ambitious set of targets agreed by members of the United Nations to eradicate poverty around the globe, but how do we know whether we are achieving them? How do we define the goalposts in order to check that we are moving in the right direction?
In order to measure the progress being made, the UN has its own expert group within its statistics division http://unstats.un.org/sdgs/ devoted to creating a framework of indicators and statistical data to monitor progress and inform policy regarding the SDGs. Key challenges include a lack of data – in some countries it has been decades since the last census, so even basic data - knowing how many people there are in each part of a country - is out of date or simply inaccurate.
Clearly, innovation in data collection and statistical techniques is needed to help improve this paucity of data. This is a key area of work for Professor Andy Tatem of the University of Southampton, who is giving this year’s RSS Beveridge Lecture: 'Mapping progress towards the Sustainable Development Goals’. In it, Andy will outline some of the key challenges, and potential solutions, to measuring the SDGs. 'Pretty much all of the SDGs are based on ensuring a certain percentage of the population has access to specific services or resources, or achieves a certain level of social, economic, or physical health’, he explains. ‘That’s why we need a regularly updated demographic evidence base, starting with the basics’.
Mapping populations without a census
In his lecture, Andy will describe some of the techniques currently being employed to measure population demographics using, for example, satellite imagery. ‘It is now possible to map the smallest of settlements and even individual buildings using automated analyses,’ he says. ‘By integrating these with small, rapid and targeted micro-censuses within statistical models, it is possible to map population distributions with appropriate confidence intervals without needing a census.’
There is also strong evidence that geographical variations in rates of poverty are linked to factors such as remoteness, urbanicity, mobility, social network characteristics and consumption - all of which can be mapped using satellite and mobile phone data. This enables further insights into a given population. 'By leveraging and integrating these multiple data sources,' he says, 'we can map key population characteristics that form the SDGs accurately.’
These techniques are serving to fill 'data gaps’ where key population data is scarce or out of date. Andy is director at WorldPop, a project which aims to improve the spatial demographic evidence base for low- and middle-income countries by integrating data from ‘traditional' sources such as censuses and surveys and matching them to digital boundary maps. 'We then develop scalable methods and models for integrating ancillary data sources to complement and fill data gaps in these traditional sources, often integrating new technologies, including high resolution satellite imagery and cellphone data,’ he explains. These are compiled into gridded demographic maps and made available (open access) through the WorldPop web portal at www.worldpop.org.
Andy initiated Worldpop after working on the Malaria Atlas Project (www.map.ox.ac.uk) at Oxford to improve the mapping of malaria risk. ‘Here we put significant effort into constructing detailed maps of the prevalence of malaria, but soon realised that we needed similarly detailed population distribution maps for malaria endemic countries to provide denominators, and that these were either lacking or outdated.’
Barriers to accessing data 'disappearing'
Most of the data used and produced on Worldpop is open data - however, accessing very high resolution satellite imagery comes with costs and access restrictions in place. Many national statistics often remain behind locked doors. And access to mobile phone operator data can be tricky due to concerns about confidentiality. This is something Andy encountered as co-director of Flowminder (www.flowminder.org), an organisation which collects and integrates anonymous mobile operator data with satellite and household survey data to analyse population movements and support strategies designed to tackle diseases like malaria, cholera and Ebola. ‘The de-identified mobile phone data we utilise within Flowminder is highly sensitive, and there are quite rightly certain barriers in place to protect customers' confidentiality,’ he says. ‘We work within international and national regulations to simply access these type of data, rather than obtain them ourselves.’
Despite these restrictions, Andy reports that accessing high quality and complete satellite imagery, GIS data and national statistics is easier now than when he first began work on WorldPop and Flowminder. ‘It is exciting to see many of these barriers disappearing, and the amount and quality of available geospatial data is rising every day,’ he says.
Andy believes that integrating multiple datasets in the ways described above can bring together the strengths of each to overcome the many weaknesses of each. 'It is likely that only through data integration will we be able to measure and map progress towards the SDGs in a regular, consistent and subnational way.’ However, he warns that new data sources and statistical techniques will continue to rely on, rather than replace, ‘traditional’ approaches such as censuses, surveys and registration systems. 'These will continue to provide the bedrock of SDG measurement and we should not neglect them in a rush towards new data sources and modelling methods.'
Andy Tatem is a professor in geography at the University of Southampton and director of the WorldPop project. His work is focused on high-resolution global mapping of population distributions and characteristics, the dispersal of diseases and their vectors through global transport networks and quantifying population movements in relation to malaria elimination planning. He pioneered the use of mobile operator data for use in public health in low- and middle-income countries in Tanzania in 2008, and is co-director of the Flowminder Foundation.
He will deliver the RSS Beveridge Lecture on 22 June at the Royal Statistical Society. Register to attend on our events page.