As part of the Bloomberg Philanthropies City Data Alliance, Public Digital will teach city leaders how to move beyond a portal-only strategy and channel data into tools tailored to the internal and external stakeholders’ needs.

The essence of this transformation is to shift cities from thinking of data as something they might publish simply for transparency, towards implementing what we call “Data as a Service”.

In this blog post, we’ll look at who decision makers and innovators are and what they need from government data and analytics services.

Government has always had a crucial role in providing data and information to various users: look at institutions like national statistics services, land registries, or national weather services, many of which have been around for decades. At a local level, city governments might provide information about local amenities, such as parks and playgrounds, and public services, such as bin collection. But the increased use of digital technologies and growing sophistication of sensors and satellites means there’s more data available now. And the importance of data as a fuel for innovation, including the development of AI, have expanded the range of users it’s of interest to.

It’s worth taking a quick diversion here to consider the distinction between data and information. For us, data is the kind of stuff you might find at the end of an API, or CSV files you might download for further analysis. Data might look like statistics, real-time sensor readings, spreadsheets, or even geographic data and satellite imagery.

Raw data can be difficult to handle due to its size and granularity, so it frequently needs to be processed to make it usable, in the same way as raw ingredients might be washed, peeled and chopped to make them easier for people to cook with. Services may create refined data from data – providing APIs over static data, statistical summaries, or projections into the future.

But even this refined data itself isn’t generally consumable by people: normal people don’t read JSON files or eat uncooked potatoes. They need information: graphs, visualisations, summaries and stories that make sense of data.

Government has typically published information, rather than data, providing its own interpretations of the raw data it holds. Continuing the food analogy, these are like government-provided meals. Part of the push towards opening up government data in the late 2000s was to enable startups and civic technologists to get access to this underlying data so that they could create their own information – different interpretations, cuts and analyses (new recipes, to continue the analogy) – to meet new and unforeseen requirements.

Now, in the 2020s, we are starting to see greater complexity and sophistication in the ways in which this underlying data might be used. Automated decision support systems, recommendation engines, and other AI systems provide an additional layer of interpretation and clarity over raw data that can be personalised and contextualised (but also biased, misleading and inequitable). A nice example of this from New York City is a project to optimise their street and sidewalk cleanliness inspection program through the digitisation of routing and navigation. These services don’t just provide information, they give advice.

So when we think about users of “Data as a Service”, it’s worth distinguishing between those interested in accessing raw data – our data caterers, who might be developers, researchers or data journalists – and those who are interested in information and advice based on that data – who might be policymakers, commuters, parents, healthcare providers or construction engineers.

It’s also worth noting that while knowledge management literature might refer to “information”, people don’t tend to call these “information services”. In our work, we will refer to services that provide information and advice as analytics services, to emphasise that these are based on and enabled by data, but form a separate layer of interpretation.

The following table highlights these different types of users and why they might want access to data and analytics.

Purpose	Data users	Analytics users
Hold city government accountable	Data journalists and civic technologists	Citizens and civil society organisations engaging in democratic processes
Make evidence-based policy	Social researchers in academia and think-tanks	Politicians and policymakers
Improve people’s lives and work	Developers in the private and third sectors, as well as city government	People and organisations, including city government and public service providers, as they live and work

For our purposes, we are most interested in city governments who want to provide data and analytics in ways that improve people’s lives and work. These are things like transport information that helps commuters route around traffic jams; information about schools that helps parents choose where to send their kids; information about utility provision or population demographics that helps businesses choose where to set up shop; or information about physical infrastructure that helps public contractors fix roads or building inspectors prioritise inspections.

The information itself might be delivered through websites, apps on our phones, public displays, or even turned into physical forms like posters and printed QR codes, depending on the target audience’s needs and capabilities. It might be embedded in and support more transactional services – people frequently need to understand the state of the world before they decide to act on it. It might be passed on through intermediaries, such as social workers, citizen advice bureaus or community groups. These are “digital services” in that a computer will be involved in generating outputs, but, when we look at it as a whole service, we can see that information might not always be best delivered directly or through digital channels to reach those who need it.

The innovators who create information from data – predominantly developers of these digital services – might be in city government, in the private sector, or part of civil society. Governments definitely can’t and shouldn’t create all the possible end-user-facing products and services themselves, just as one restaurant couldn’t possibly supply the variety of meals we might want to eat, so there are interesting choices for cities to make about where it draws the line.

For example, should it develop an app itself, potentially competing with the private sector who could provide a better one? How should it advance equity and ensure that marginalised communities, including those who are digitally excluded, have access to information, if neither the market nor local civic technologies cater for them? How should it take active steps to fill the gaps it can see, and that meet its own priorities, while also removing itself as a blocker to unexpected, serendipitous reuse of data by others? The answers to these will vary depending on the vision the city has of itself and its role.

When data services are working well, they become infrastructure that underpins a range of other services and activities. We become reliant on them, just as we do on the plumbing in our homes and offices. Unfortunately, infrastructure is boring until it breaks, which can lead to a focus on the more visible, and flashy, analytics services at the expense of the data services that make them possible.

Where city governments do provide data, we think that continuity in the supply of that data is essential. Just as we can’t cook healthy meals with out-of-date ingredients, and restaurants can’t offer a standard menu if they have an unpredictable supply, all the products and services built with data require that data to be provided in a timely, predictable, and reliable way. As cities have recognised this, there has been a shift in emphasis towards “fixing the plumbing”: focusing on consistency and sustainability of data publishing to unlock the ability of others to build a range of analytics services using them.

We believe both are important. Ensuring a steady, sustainable, reliable supply of data is a necessary step towards providing the analytics services that people need. But it isn’t sufficient. The analytics services need to be built too, either by the city itself or by third parties. Building analytics services in parallel with data services helps to both reveal requirements on data services, making them more useful, and motivate and inspire further uses of data that makes the investment in it worthwhile. Cities need to build both.