Spreading the word - data science, infections and mobile data at Scotland’s Data Summit
Academics are supposed to be mind-expanding, challenging and out-there. So when you see one talking about using a technique called body-shocking, giving machines spending power, and swapping Pokémons for Bitcoin in a virtual collecting experiment, it shouldn’t be surprising. Yet somehow it is and in the best possible way.
Professor Chris Speed, chair of design informatics at University of Edinburgh, certainly conformed to type in his presentation on day two of Data Summit 18, organised by The Data Lab as part of DataFest18 in Edinburgh last month. But his job is to pay ideas forward based on stretching what is possible now.
His experiments are designed to explore what happens in “the near future when we start to interact with objects that have been given agency,” he explained. “Organisations are already using, storing and selling data, much of which has been created by bots which changes our assumptions about how you create value.”
“Does giving a thing agency change our relationship with it?”
In one example, his research team worked on what happens if the internet of things has its own money. “Does that give a thing agency and change our relationship with it? You can devise one-millionth of a Bitcoin and give anything a wallet, but what would it do with it? It opens the possibility of smart contracts between things and humans which is hard with the old concept of money,” he said.
Using the technique of body-shocking, he sent researchers out into the physical world to carry out virtual tasks in which fractions of Bitcoin could be earned by travelling to certain locations, then exchanged for goods such as cinema tickets. The video of their efforts was a conference highlight, but in equal parts baffling as it was informative. As Speed’s own wife asked when he showed it to her, “where’s the cinema?”
That could become a meme as we learn how to blend human behaviour with newly-enabled technology. When Speed put out coffee cups at a conference that could be charged with value that was exchangeable for a free drink, but only if delegates engaged in conversation with another delegate, they proved to be a hit. But perhaps only because the control cell was a surly Danish barista who also gave out free coffee, but without speaking to his customers.
“With cash, you don’t know where it’s been. With Bitcoin, you do.”
At least one purpose of such experiments is to explore the boundaries between humans and things in order to understand how our adoption is affected. As Speed noted, “it’s about adding values to value. With cash, you don’t know where it’s been. With Bitcoin, you do.” Bearing in mind that 11% of UK banknotes have been found to carry traces of cocaine, that could be an important factor in any transition to a cashless society.
That fact in itself demonstrates the way human behaviour can spread all kinds of matter through physical contact (or just proximity), something which epidemiologists are increasingly turning to data to help them to understand. Hannah Fry gave the conference a powerful insight into this kind of work with a pre-broadcast look at extracts from her experiment, “Contagion! The BBC Four pandemic” (available on the BBC iPlayer until 21st April).
The biggest citizen science experiment of its kind, it invited individuals to download an app which tracked their location hourly for a 24-hour period. All data collected is sent to a research team from the University of Cambridge and the London School of Hygiene and Tropical Medicine.
Data collected between September and November 2017 was used to simulate the spread of a highly-infectious flu across the UK and predict how many of us might die, with the results revealed in the programme. Data collected between December 2017 and December 2018 will contribute to a new gold-standard data set for use in future simulations and in wider pandemic research.
In her preview, Fry showed herself acting as Patient Zero in one small village, casually infecting residents and then using tracking data to show how that infection spread. (The infection is actually just the app registering the proximity of another mobile phone running the same app.) The bad news is that within some 100 days virtually the whole UK population would be at risk. The good news, she was able to exclusively reveal to delegates, is that the Hebrides would escape the outbreak. Stories of packed trains leaving for the isles from Waverley Station that afternoon may or may not have been fake news.
“We are trying to answer how to unlock data in places that are hard to get to.”
Mobile phones and the data they generate were a recurring theme across the conference - hardly surprising given their centrality in the lives of consumers. A notable strand was the potential for using mobile data for social good. Stefaan Verhulst, co-founder and chief researcher at Gov.Lab, noted that, “we are trying to answer how to unlock data in places that are hard to get to,” citing the earthquake in Nepal as a prime example. Ncell, the regional telco, shared its data with a Swedish data science firm to track population movements around the affected areas. “That had a crucial impact for aid organisations and government response,” he explained.
As valuable as such efforts may be, they are not always easy to bring about. “Companies are not willing to share because data is a core asset,” noted Verhulst. That is where Gov.Lab can step in to act as a neutral stakeholder which is able to bring together various agents and data owners to resolve problems, using one of its six different frameworks.
“Not-for-profits don’t have the resources of big companies like Netflix.”
Similar work has been done by DataKind which pulls on the altruism of the data science community to transfer skills into the not-for-profit arena. As its lead data scientist Erin Akred said: “DataKind exists to use data science to create the world we all want to see. Not-for-profits don’t have the resources of big companies like Netflix.”
Some times this works the other way around, with large organisations volunteering to help out on world problems without the prompting of a third party. Shakeel Khan, data science capability building manager at HMRC, spoke about a project he engaged with in Pakistan. “They have a severe problem with Dengue fever. Using mobile phone data, we have been able to track its spread and stop it at source. That is saving lives,” he said. The mosquito-borne disease can be reduced or prevented through a range of measures, from sleeping under mosquito nets through to emerging vaccines, provided these are deployed in the right place.
The spread of data science can be seen as having its own epidemilogy with outbreaks in clusters that then affect local populations. Khan reported on his visit to Rwanda which is fuelling its progressive culture and addressing economic growth in part through a data modernisation project which includes the creation of a 200-strong data science centre. HMRC itself, meanwhile, is using new techniques to tackle non-compliance and fraud which have a £29 billion potential payback.
It is these kind of unexpected and forward-looking insights that make conferences worth visiting and from this perspective Data Summit did not disappoint. In fact, DataFest 18 as a whole revealed just how rapidly Scotland’s data community is growing and the economic potential which is likely to result, with some 3,000 visitors participating in 45 events across the week. Like the atmosphere among delegates, that kind of enthusiasm and growth is infectious.