State of City Data: Data and Vulnerability

In the first issue of The State of City Data, Hillary Simpson explains why joined-up data sharing is essential for protecting vulnerable people in our cities. And why the only reason to collect public information is to act on it.

I’m interested in the human experience of living in cities. For me cities are only smart, welcoming and functional if they work for us all – including outliers such as vulnerable people. I’ve worked for many years in front line housing support services. More recently I’ve worked to join up personal data silos to safeguard children.

The two particular issues I’m passionate about are domestic violence and child protection. I measure whether a city is smart depending on whether women and children are safe. This is difficult to see because these issues happen behind closed doors.

In 2017 the Police Foundation reported:

‘The police service and community partners are facing an unprecedented increase in local demand generated by complex social problems such as mental health related incidents, missing children, homelessness and substance misuse. Alongside this, we have seen a greater willingness to report previously ‘hidden crimes’ such as child sexual abuse and domestic violence, which are more complex to prevent and investigate.’

City data sharing can play a big part in helping to ensure vulnerable people are safer.

A wider context

I’ve travelled to a number of cities in the past year while working on city data projects. I’ve tried to find quantitative data on the prevalence of domestic abuse to investigate trends and mapping.

We already know domestic abuse is a very expensive problem: the LGA reports that domestic abuse costs public services across England and Wales over £3.85 billion per year. Living Without Abuse report that a single domestic violence death costs public services £1 million. This doesn’t include the devastating effects on subsequent generations. Women’s Aid is trying to help reduce long-term effects by empowering survivors through their work.

But are we giving sufficient priority to vulnerable people in our open data sets? It’s important to uncover ‘hidden problems’ to see how data can be used to tackle them. When complex interrelationships between people break down, we all pay.

Information sharing for serious incidents

Better use of data may have resulted in better responses for some of the most serious public sector incidents. For example, the Independent Grenfell Tower Recovery Taskforce Initial Report found that:

‘…it is hard to understand why the various responders continue to say they don’t have a common and comprehensive list of survivors and displaced residents, where they are currently living, and what their assessed needs are…..Various lists do exist….rapidly sharing this data with those that need it to deliver services, should not be beyond the technology that is readily available.’

For over ten years, child protection serious case reviews have reported that the lack of information sharing is a factor in child deaths. The Baby Peter Connelly (Baby P) case and Lord Laming’s 2003 report on the Victoria Climbie case demonstrate this.

In the case of Baby P, one agency knew that an additional adult male had moved into the household but this wasn’t shared with the social worker. This meant the increased risk wasn’t taken into account.

What’s stopping us?

Firstly, there may be a perception that sharing personal data is illegal or unethical. Of course we need to respect privacy – and the General Data Protection Regulation (GDPR) helpfully sets standards here, using Data Protection Impact Assessments.

This is a balanced risk assessment approach – there is a real-world impact on vulnerable people of not sharing personal information. Sharing limited demographic data and indicators (not case data) has proven to save money and lives. And if it’s done at a local (rather than national) level this is what citizens expect.

Secondly, there is a perception that unless data strictly adheres to standards and is clean and cross referenced (NHS numbers for example) it cannot be accurately matched. The latest machine learning technology gets over this issue. Tuned by large data sets and checked by the human eye, large volumes of data can now be matched in realtime.

The London boroughs of Brent and Camden have been linking up their legacy system database silos for years. This technology is now readily available through a range of low-cost SMEs. Every city can build on the work of pioneers without having to start from scratch.

Projects that work

A range of local authorities, cross agency consortia, professional bodies and SMEs are showing the way by carefully joining up personal data silos and tackling some wicked issues.

  • The Low Income Londoners project tracks households over time to tell policymakers what support is effective. The evidence base they’ve developed is a powerful tool capable of influencing central government
  • The CIPFA Counter Fraud Centre helped local authorities save an estimated £271 million in 2015/16. But more can be done by individual authorities
  • In the USA, an organization called Thorn use data to identify victimised children and act

And since 2016 Spotlight, with their cutting-edge technology, has helped law enforcement:

  • Identify an average of 8 children per day
  • Reduce critical search times by 65%
  • Find a total of 5,791 victimized children

A number of agencies in the UK want something similar, according to the ‘Time to listen’ report (2016).

Why now?

The only reason to collect public information is to act on it.

In order to help with early interventions and prevent costly duplication of services there’s no avoiding linking up silos of personal data within public services. The examples above need wider and faster uptake in order to save lives and money.

Thanks to GDPR, our public sector leadership teams have never had a greater understanding of information sharing.

There are a range of low-cost machine learning tools now available to accurately link up personal data between individual line of business databases. This works even if front-line staff haven’t rigorously followed data standards or have mistyped in a hurry.

The focus for city data needs to move on from quantifying the hard assets – the internet of things – to helping vulnerable people within our cities, especially as we move towards smarter cities of the future.

Hilary Simpson is the founder of Sleuth Co-op, and previously worked with Nesta on their Office of Data Analytics programme, where she had been seconded from Camden City Council having led the development of their Camden Residents Index (CRI): a Master Data Management solution linking demographic people data from 16 line of business systems.

The State of City Data is a series of thought leadership pieces from leading data practitioners who give their assessment of the current state of city data. These pieces form a comprehensive report of what works and what isn’t working for data in our cities. And what examples should be scaled and where we can learn from heroic failures.