Cloudy Futures

Claire is a Senior Product Manager at Arup

On 8 June, I was a panelist at CogX2020 — a conference that is self-described as a “Global Leadership Summit and Festival of AI & Emerging Technology.” This year the 30,000+ person summit went ahead as a completely virtual conference. The theme of the conference was: Getting The Next Ten Years Right.

At CogX, I joined a panel to discuss what happens when the data for planning, operating, and organising our cities is all uploading into the cloud. My fellow-panelists were representatives from Google (Nick Taylor, Migration Practice Lead) and Darktrace (Andrew Tsonchev, Director of Technology). We were led in conversation by Andrew Eland (Founder of Diagonal, and collaborator of Arup’s City Modelling Lab — among other work.) The title of the session was:

Cloudy future: What are the risks and rewards of uploading a city into the cloud?

There are a multitude of considerations packed into the potential consequences of hosting all data about the public, for the public, in the cloud. I focused on my experiences working with Arup’s City Modelling Lab and Advanced Digital Engineering Software team to talk about the ways I understand cities to use data.

My work is motivated by a belief that we can do better when it comes to imagining and building transparent, open, and inclusive data systems. It was a delight to use the CogX2020 stage to highlight some of the societal and policy challenges of working with data.

Q: What kinds of data are cities working with, what kind of problems are they trying to solve?

A: My work focuses on data related to city planning, and transportation within cities. Transportation data is a great microcosm of data systems in cities. Transportation is fundamentally about connecting people to all the things that we do: live, work, learn, seek health care, play, and shop.

There are two halves to understanding, planning and operating transportation: transportation supply tells us about what infrastructure and services people can use to move around; transportation demand tells about who is making a trip, when, and by what mode (walking, cycling, driving, taking a train, etc.) Data within these two halves is heterogeneous. Some data is public (e.g. bus schedules, road networks, census data). Some data is private (e.g. private transportation services like scooters, ride-share apps, or taxi fleets; GPS data traces from mobile apps.)

Understanding transportation systems, and making use of data that represents transport systems, can be useful in achieving the outcomes cities strive for: safe, healthy, prosperous, low/no-carbon communities. For example, we know that the time that it takes someone to access education, health services, or work directly impacts the quality of education, health outcomes, and income of that person (NatCen, Transport and inequality: An evidence review for the Department for Transport. 2019.)

However data has a real cost. The costs of collecting, governing, managing, analysing data can stack up. The benefits of good data are often shared across multiple stakeholders, public and private. When costs are incurred in one place but the returns from that investment are found in many place s— it can be challenging for any one actor to justify the investment of improving data collection or systems.

As a result, we often see people trying to reconstitute data to serve purposes it was not intentionally designed for. For example: in London, Oyster card data allows TfL to ensure that passengers have paid their appropriate fare for their ride on a TfL bus or the underground. But this data also tell TfL a lot about when the Underground is busiest. This information allows TfL to fluctuate fares and incentivise people to travel when the underground is less busy. But if TfL want to understand which lines are used most, Oyster data can only hint at this information. In order to answer this question, TfL either needs to collect a new data type (which they started doing in 2019 with WiFi data) or make assumptions about the relative popularity of underground lines.

The point here is, it is important for policy makers to be motivated and lead with vision when it comes to designing the data they need to tackle city challenges. This is important because knowledge is not emergent. Knowledge doesn’t just come from the data. Knowledge has to be hypothesis driven, and a hypothesis should be tested or validated with fit-for-purpose data. Our current, heterogeneous state of data means that our city leaders have a lumpy view of our cities.

Q: The theme of CogX is “Getting the next 10 years right.” What needs to happen over the next ten years to ‘get it right’?

A: I’d like to see a future with greater transparency of accountability in our data systems, healthier competition between service providers, and more power for individuals to design collective data systems (including surveillance systems.)

Ownership and accountability

We need a wilder imagination as we think about what is possible when it comes to data ownership, and the expectations we have of public institutions. I am thinking of collective ownership, and broad, inclusive oversight of data use. I believe data about services (e.g. electricity use or bus services) and data about people (e.g. employment figures or heath information broken down by sex or race) can be valuable. Communities should be the beneficiaries of their own, collective data. This is a complex dynamic, and something that can easily slide into surveillance if we can’t imagine and design new ways to share ownership over data systems. What might this look like? As Bianca Wylie has demonstrated, we don’t have to have all the answers. We can hold space by asking questions and seeking accountability from the institutions that should work for the public.

Healthier competition

As we build on existing- and brand new data services, products, and even specialist analyst capabilities — how do we keep competition? Competition is important to mitigate risks associated with vendor lock-in. It is also important to allow for innovation and more imaginative (dare I say, “disruptive”) options. Interoperability (the ability to move data between systems) is one part of allowing for competition.

We have seen the benefits of common data standards: The General Transif Feed Specification allows apps like Google Maps and City Mapper to make use of public transport schedules. Open banking lowers the cost of customers switching banks, among other benefits. But what happens if a company, or public agency, wants to switch cloud providers?

The next frontier for data standardisation and interoperability is in cloud infrastructure. How do we ensure it is as easy to switch between cloud providers, as it is to switch bank accounts? And how do we encourage healthy competition in this meanwhile-time? Companies and public agencies may consider building a a diverse portfolio of tooling across different cloud providers.

Surveillance — what’s the exit strategy?

Surveillance technologies have been increasingly common. The recent track-and-trace apps being used or considered to assess the risk of covid-19 to public heath — these have put global attention on the lack of safeguards we have when it comes to surveillance technologies. Our governance systems are behind our technology systems. This means that we don’t have clear consequences for breaking agreed terms of use. I’d like to see more transparent definitions for how data can be used, processes for how people can participate in those definitions, and the consequences for misuse.

For more highlights from CogX you can visit their YouTube channel.