An open letter to the next Mayor of London explaining why they need a London Office of Data Analytics.
To the new Mayor of London,
Congratulations on your election victory! No doubt you’re impatient to start delivering on your campaign promises. Just before you do, I’d like to highlight one point: no matter what your policy priorities for London, you’ll need data to deliver them.
If you want to tackle urban deprivation, improve air quality, combat congestion or increase the availability of affordable housing, you’ll need data. Data to diagnose the scale and distribution of those problems and the demand to be met.
If you want to support London’s boroughs and public sector bodies in reforming public services, both you and they will need data. Data that spans the whole of the capital to see how specific issues transcend local authority boundaries. Data to marshal the city’s finite resources at areas of greatest need. Data to predict and prevent problems from occurring – or at least intervene earlier when they are simpler and cheaper to resolve.
Did you mention that you wanted to support London’s businesses? You’ll need data for that, too. Data on how transport routes affect economic activity, data on consumer spending by street, data on footfall statistics. And would you like to model the economic impact of your policies before you implement them? You can! But – well, you can see where this is going.
In short, though in isolation it’s no panacea, it’s hard to overstate the importance of data to the success of your agenda.
So here’s the problem.
When you enter City Hall for the first time as Mayor of London, you will find that the data at your and your staff’s disposal is very limited.
Let me be more specific:
– With the exception of planning applications, City Hall does not systematically collect any data from the 32 London boroughs or the City of London.
– Most of City Hall’s data comes from central government departments, but that data is often already aggregated, preventing granular analysis.
– With just one exception (Transport for London’s Unified API), the data is historic and static, not live and dynamic.
– The datasets you do have access to are only available because someone at City Hall thought to ask for them. There may be thousands of other datasets held by central government departments and other agencies that would be useful to you but about which you have no knowledge.
– City Hall has around 40 staff in data-specific roles, doing excellent work in the field of data science and analytics. Yet the majority of their time is spent producing reports, funded by one-off innovation grants, that provide a moment-in-time snapshot on specific issues. Useful, absolutely. But they lack the continuous funding and resources to provide you with live data insights to inform your strategic and day-to-day decision making.
It’s not just City Hall that faces data challenges. Consider London’s boroughs.
All of them use and create data in the course of running their services. Some have grown very sophisticated at mapping their data to show the location and intensity of the issues they address. But, for the most part, they don’t have access to data on those exact same things beyond their boundaries. Like a jigsaw that has never been put together, London’s local authorities have all the pieces, but no-one – not even you as Mayor – can see the whole picture.
Why is this? Different organisations use different IT systems, some of which don’t easily enable data sharing. The public sector uses dozens of different data standards, formats and conventions, creating the digital equivalent of comparing apples and oranges. And there are cultural, organisational and (perceived and real) legal barriers: organisations feel protective of their data and are, rightly, cautious about sharing it.
Whatever the causes, this data blindness is a serious problem given that Londoners don’t conveniently live and work out their lives in one borough. Communities, business districts, areas of deprivation, crime, littering and school catchment areas frequently cut across borders. London cannot live up to the challenges it faces as 33 separate islands.
And despite being widely regarded as a world leader in the field, the open data available to London’s citizens and businesses has several shortcomings.
First, it’s often released at too small a scale to enable significant innovation. Less than a third of London’s boroughs share their data via the London DataStore – the capital’s official open data portal. The rest release theirs on their own websites or not at all. This fragmentation makes it incredibly hard for potential developers to create a viable business model from an open data product. When the typical price of an app is around 69p, and revenues from advertising are just a few pence per click, developers need a larger potential customer base than the residents of just one local authority area. (And who wants 33 separate apps just to park or report potholes?)
Second, few investors and entrepreneurs will risk their own time and money to build products with open data unless its provision and quality are guaranteed. But no such guarantees can be made. That’s because releasing open data is not cost-free. At the very least it requires the resource time to provide sufficient support for those who wish to use it. Given the current pressure on public sector finances, some councils may cease providing open data altogether if they see it as nothing but a cost centre.
Third, the flow of information is in just one direction. Businesses, charities and citizens can receive data from government. But there is no official mechanism to provide data to government. That’s a huge missed opportunity when many – if not most – of the highest value datasets about the city are held by corporates, charities and citizens.
For one of the world’s leading global cities, I don’t believe this is adequate.
If you agree, I have a suggestion for what you could do.
We need a London Office of Data Analytics (LODA).
The model is inspired by the Mayor’s Office of Data Analytics in New York City (described here), but significantly adapted to meet London’s particular needs and very different political environment.
LODA would be a small team, based in City Hall, that had the resources, technology and expertise to bring together, translate, analyse and provide actionable insights from data sourced from all local authority and public sector organisations across the capital. Those insights would be made available to improve data-driven decision making by the 32 boroughs and the City of London, London’s public sector bodies, the GLA, and, of course, you as Mayor. A subset of LODA’s data would be made available to the general public as open data via the London DataStore, which would be converted into a city data market that connected organisations and individuals that had useful data with those that wanted it.
LODA would be led by a Chief Data Officer (to be clear, a separate role from the Chief Digital Officer proposed by the Centre for London). The team would be made up of three parts – a model that Manchester is currently pioneering with its GM-Connect programme:
1) Data Analytics team – responsible for using data analytics to: combine datasets from different sources to create London-wide maps showing how specific issues / demand / events / objects transcend borough boundaries; target resources at areas of greatest need; spot correlations that can predict where problems will occur to enable preventative action; model the impact of emergency scenarios / future growth / new legislation; create visualisations so that data can be understood by all policymakers; provide open datasets that cover the whole of London.
2) A technical team – responsible for implementing and managing the technology required to automatically pull data from IT systems in London’s public sector bodies and translate between the different data standards they used.
3) A legal team – responsible for providing hands-on Information Governance, legal advice and contract-writing assistance to ensure that public sector bodies are able to share their data with LODA responsibly, ethically and in accordance with all relevant data legislation.
LODA’s creation need not be expensive.
The New York version started with just two people using Excel spreadsheets. They only expanded their operations where they could prove that data interventions provided a clear return on investment. In London City Hall there are already staff with the requisite data skills. And thanks to Witan – a project funded by Innovate UK in which the GLA is working with an SME called Mastodon C to build a city modelling platform – you have much of the technology you need as well.
What would such a team enable? Almost anything you can imagine, but to give a few examples:
Support the expansion of shared services. By creating maps of demand that span the whole capital, LODA could help councils make more informed decisions about where they can join with neighbouring boroughs to run or jointly fund specific services – a proven means of making significant savings.
Prediction and prevention. By correlating datasets from different sources, LODA would be able to help predict where problems are likely to occur in the future and support public sector bodies in targeting early interventions more effectively. This could be applied to anything from tackling the blight of beds-in-sheds to improving support for troubled families; and from fighting tax and benefits fraud to improving the effectiveness of food safety inspections. (This would be achieved using the 10 step data model developed in New York City.)
Support business growth. by using the city data market to source corporate datasets, LODA could help businesses find the best location to set up shop and optimise opening hours. For example, mobile phone operators collect data on the location of their users. That data could be aggregated and anonymised to show footfall in every street at different times of day, or days of the week. And with 75% of all spending in the UK retail sector made using credit and debit cards, companies like MasterCard have data on patterns of consumer behaviour within a city, which could reveal where certain types of products are favoured.
Spread digital skills. LODA would become a catalyst for promoting and extending the use of data analytics throughout the city’s public sector bodies. Having created a data model to improve a particular service (as described here), they would train up and delegate its running to the team responsible for that service. LODA could also collaborate with local universities and the Digital and Future Cities Catapults to run data analytics training courses for public sector leaders and workers.
Improve Open Data. LODA would solve the three shortcomings of London’s open data. By enabling public sector bodies to use their own data to drive service efficiencies, it would make open data financially sustainable for the long term and improve its quality. By stitching together datasets from multiple sources it would provide data covering the whole city. And by converting the London DataStore into a City Data Market – it could help open up and access data from businesses, charities and citizens, too. (This is not fanciful: the GLA’s Andrew Collinge is already leading the way with the recommendation to create such a market or “City Data Exchange” in his City Data Strategy.)
For this to work in London, you would need to do three things:
1) Make a public, long-term commitment to data-driven government. Data will only deliver genuine transformation to the way London is run if the message is delivered right from the top that this is how London’s public sector works. It will take your leadership to overcome the cultural and organisational resistance to change.
2) Appoint a Chief Data Officer. They must be a senior mayoral adviser responsible for ensuring data is completely ingrained into decision making at City Hall.
3) Work to win the backing of the boroughs. Much of the data on which LODA would depend would come from London’s boroughs. Securing their support needn’t be impossible – LODA is neither a partisan idea nor a tool to increase the control of the Mayor’s Office and the GLA – it’s a fundamental tool of good government. But pointing out two things would help win their backing:
First, don’t call it the “Mayor’s Office of Data Analytics” (as per New York), but the London Office of Data Analytics. This is not just semantics. It highlights that LODA is a resource designed to benefit the whole of London, not just you as mayor. In fact, perhaps it deserves a better name altogether – how about the “London Centre of Data Innovation” (proposed by Andrew Collinge), “London Data Insights Team”, or “London Data Service” – that emphasises making the data useful and used rather than being an end in itself.
Second, let it start by focusing on data initiatives that support the boroughs – particularly their urgent need to reform local public services. This is something they can’t do for themselves. If each borough tried to negotiate individually with the other 32 councils to share their data, it would require setting up 528 one-to-one data connections. By contrast, LODA could set up a single data exchange with each council (33 in total) to bring their data together in one secure location. This would save a huge amount of time, money and effort – not least in translating between all the different data standards used across the capital. (And as Paul Maltby, Director of Data at Government Digital Service, has suggested, even central government struggles to hire experienced data scientists.) Additionally, LODA could combine local authority data with datasets from the Metropolitan Police, the London Fire Brigade and other London public sector organisations to provide insights in a way that would not be possible on a borough-by-borough basis.
So where now?
As things stand, as Mayor of London you do not have visibility of what is happening in the city you have been elected to lead. London’s councils and public sector bodies do not have the data they need to deliver reform at a scale sufficient to meet the daunting financial challenges they face. And London’s businesses, charities and citizens do not have the open data they require to support significant innovation from outside of government.
A London Office of Data Analytics is the first step to address all these shortcomings.
It will take your leadership to establish.
And if you’re looking for an organisation that has the time, resources and expertise to make it happen, Nesta is here to help.
Director of Government Innovation
Nesta Innovation Lab