Powering up data for future public services

Data is the foundation for the next generation of public sector services. Civica’s Deane Greenouff discusses how to make the most of this valuable asset

For public sector organisations today, data has become a highly valuable commodity. According to the Centre for Data, Ethics and Innovation (CDEI), it’s “a powerful asset for local authorities”, helping them to make better decisions, allocate resources more efficiently and create greater transparency for citizens. The CDEI report cites the recent pandemic as having given a huge boost to the use of data, for example using information at local authority level to identify vulnerable residents or using population-level data to contain local outbreaks.

AI changes the game

More recently, councils have been adopting machine learning (ML) and exploring artificial intelligence (AI) to improve and streamline how services are delivered. Already, a huge range of potential applications for this high-profile technology are being put to work. AI is being used to review and help manage planning applications, collect business rates more effectively, take 99% of the effort out of analysing clinical data, and to make low-cost preventive maintenance possible using internet of things (IOT) connected devices.

Data friction

Large data sets can be hard to manage if not properly connected and organised. Organisations of all sizes will be familiar with the huge headache of working with data that’s held in different forms across different systems, for example on a mix of spreadsheets, paper documents and legacy on-premise systems. Our research of public sector professionals shows that 85% believe missing, incomplete or inaccurate data is negatively impacting customer experience. The administrative friction this creates has driven the trend towards migrating onto cloud systems – a move that requires clean and well-organised data to be successful. Effective master data management is the answer – it can be used to constantly cleanse and maintain data to ensure disparate systems can be joined up to provide a single view of the citizen.

Beware of bias

It’s worth noting that AI and ML models can sometimes pick up on imperfections and biases in data sets, creating outcomes that are sometimes the opposite of what the service provider wanted. When exam boards had to predict grades based on past performance, for example, a predictive analytics model downgraded high-performing students from schools that had performed poorly, because those students tended not to do well. And in the Netherlands, an algorithm designed to identify benefits fraud was found to discriminate based on race and gender. Importantly, it did this even though it had deliberately not been given applicants’ race or gender. Instead, it had used other data, such as Dutch language proficiency, as a proxy.

Stories like these can make people nervous about sharing their data, just at a time when public sector organisations have an opportunity to build valuable new services on it. However, a recent Boston Consulting Group report shows that people largely support sharing their data to improve services, especially with the NHS. So, part of any successful data-driven strategy will involve being transparent about how data is used, and for what purposes.

Where to focus?

So what can local authorities and other public sector organisations do to ensure that their data is more of an asset?

1. Fill the skills gap: we need to equip leaders and service owners with broader data skills. Managing data and its interaction with algorithmic models and AI is a specialised task. Given the value of data-driven automation – and the potential consequences if it goes wrong – public sector organisations need to attract the talent required to manage it effectively.
2. Communicate well: ensure stakeholders are aware of how their data is being used and for what purpose. The Unlocking the Value of Healthcare Data report recommends a specific communications campaign that clearly explains how data will be used and by whom. The report also recommends emphasising the move from data sharing to ‘data access’, a model where data is hosted on secure platforms rather than transferred to external parties.
3. Ensure data is accurate, complete, deduplicated and current: this can be a significant task – but a ‘data cleanse’ is the first step in getting to a single, trusted repository of data that can be managed easily and checked for appropriateness and bias. For example, we partnered with Police Scotland on a major data programme to integrate data from multiple sources, remove duplication of records, provide enhanced data analysis and present a single view of a person or location, while complying with all legal and ethical considerations and requirements. This high-quality data is better supporting decision-making by police officers and staff.

These focus areas can help make data intro a true asset – and the solid foundation for more efficient policy decision making and improved data-driven public services.

Deane Greenouff is Executive Director, Local Government at Civica