Back to top

How is big data providing better insights into public welfare?

Author: Kendra Schreiner
Published Date: 23 April 2018

The gap in information

Improving public welfare requires an in-depth understanding of needs and issues among different groups in society. Some mechanisms exist already to measure this information, such as the Genuine Progress Indicator, the Human Development Index, and the World Happiness Report. However, all of these are based on various datasets such as household consumption surveys, nutrition, health indicators, and more, which present a number of challenges. There may be systematic bias in who is surveyed and survey design can restrict the range of possible responses; financial and health data may be inaccurately reported or infrequently collected; and data is not real-time, with surveys often being completed every few years. Often, these studies produce aggregate figures that obscure local realities. This leaves large gaps in governments’ abilities to respond effectively to social needs.

How can big data help?

Big data can be utilised to bridge these gaps by monitoring and harnessing public opinion and wellbeing data in more innovative ways. Increased access and better insight to the changing perspectives and welfare needs of the population could help steer away from aggregate, dated measures and lead to a more holistic understanding of what’s needed  for social progress.

Amplifying societal concerns and needs. A number of cross-sector collaborations are working toward magnifying the public voice. UN Global Pulse Kampala is working with the UN and the Government of Uganda to apply AI and machine learning to analyse public radio content for insights into public opinion. With a refugee population of over 1.2 million in the country, real-time information is needed to understand what is happening on the ground, gauge public sentiment, feed early warning systems, and monitor the implementation of projects and programmes. Radio content analysis in 2016 highlighted concerns related to refugee rights to agricultural land and implications on forest usage and also revealed discussions and rumours around a cholera outbreak in refugee settlements two weeks before the outbreak was officially announced by the government. An evaluation report found that this radio-content analysis “may be used to close an information gap across the digital divide by generating machine-readable big data about refugees in rural areas where up-to-date information is seldom available.” This can then be used by humanitarian organisations and government institutions to monitor conditions on the ground and inform decision-making processes.

Social media is also a powerful tool with which to monitor public sentiment and wellbeing. The Laboratory for Social Machines at the MIT Media Lab has partnered with Twitter to use machine learning and natural language processing in order to map and analyse social systems, such as political groupings and the isolation of different social groups. The Laboratory “seeks to enable human-machine collaborations that enhance our ability to listen, learn, and engage across communities.”  A pilot project in Jun, Spain is assessing the town’s use of Twitter as its principal medium for communication between residents and the government. They are analysing the Twitter data alongside other town records to determine if public engagement has risen as a result, if the demographic composition of dialogue changing, if public issues are solved more efficiently, if citizens vote more, and if there has been a fundamental shift in the ways of governing. While the study is still ongoing, the Laboratory hopes to determine the possibility for replicability and how the system can be used to amplify citizens’ democratic voice in Jun and globally.

Understanding integration and cohesion. The European Commission is currently sponsoring the Data Challenge on Integration of Migrants in Cities, which makes 2011 Census data available that shows the concentration of migrants in cities in EU member states. Researchers are encouraged to combine and overlay this with other datasets to better understand urban diversity, spatial segregation, access to public services and housing, income, and electoral outcomes. Combination with real-time data could provide an important lens into refugee wellbeing in Europe.

Preventing criminal recidivism requires community supports over the long term. Khulisa, a non-profit in London, provides a short rehabilitative program with young offenders. However, for both them and the government it is hard to monitor ex-offenders and their well-being upon release. As such, Khulisa is partnering with community organisations in London to collect data on the integration progress of past program participants. This data is then analysed using wellbeing indicator tools, which monitor feelings such as usefulness, value, anxiety, and integration, and is compared to ex-offenders who have not received the same rehabilitative programming. Preliminary results have been positive, demonstrating the ways in which Khulisa’s programming is helping participants live stable lives. This data project is not only to drive evidence-based programming, but is being used to influence policy makers and other organisations to make well-being the centre of their approaches.

Democratising data input. Using data to better  focus on measuring wellbeing is not a one-way relationship. Some initiatives  are giving data and tools into public hands to amplify their voices and needs.

In Australia, Aboriginal peoples face lower health and socioeconomic indicators than the majority population. The South Australia Health and Medical Research Unit, supported by the Fay Fuller Foundation, is working to create a data warehouse on a range of health and cultural indicators in 18 geographical areas across South Australia by and for Aboriginal people. This project is unique as it is designed to provide Aboriginal communities with their own data, allowing them to drive how they want to engage with service providers and monitor performance of government systems on their health and social outcomes over time.  

Emerging research into and the development of smart cities also demonstrates promising ways the link between citizen opinion and wellbeing and government policy and measurement can be increased. These projects both use data mining and provide platforms through which citizens can shape planning. MIT’s Senseable City Labs uses a range of tools - such as mobile data and censors - to create a real-time city. Their “Friendly Cities” project uses call detail records to identify places in the city that bring people together, and research on “The Urban Village” used mobile communication data to study social networks in large urban areas. Sidewalk Labs’ proposal for Toronto includes a digital infrastructure that allows the public to collaborate to address local challenges. These projects highlight  innovative methods by which data can be harnessed to improve the liveability of cities and the wellbeing of residents. However, it is important to note that Sidewalk Labs is a private company rather than the government, and understanding how this relationship will work in the future is at the forefront of many citizen’s minds.  

Aggregate, inaccurate, and time-lagged data often leaves many governments blind to many important aspects of wellbeing, limiting policy effectiveness. Big data, harnessed by innovative organisations and collaboratives, is providing windows into the on-the-ground lives and needs of everyday citizens. Scaling such initiatives could help drive a paradigm shift on what counts as progress and change the focus of political agendas.

This is part of the work SIX is undertaking with four foundations from around the world to explore how data can be used to help cross-sector partnerships address complex problems. We seek to highlight successful global examples to inspire others, curate an action-led dialogue, and explore the role and entry points for philanthropies to engage with and enable data-driven ecosystems and systemic change around social challenges.