With International Women’s Day comes the regular question, why aren’t there more women studying science, technology, engineering and maths (STEM)? Part of the answer is a historical absence of role models. But anybody wanting to hear a true pioneer who is innovating with big data analytics - and setting an example for all female STEM wannabes - should make a point of sitting in on a presentation by Nuria Oliver, director of data science research, Vodafone.
With a PhD from the Media Lab at MIT, she is the first female Spanish computer scientist to be named an ACM Distinguished Scientist. She is a Fellow of the European Association of Artificial Intelligence and one of the most cited female computer scientists in Spain, her research having been cited by more than 11,000 publications. The 23rd March sees her giving a keynote presentation at DataFest 2017 in Edinburgh, where she will be revealling some of her recipes for success with data.
“My area of expertise is modelling human behaviour through data,” she told DataIQ in an interview. “How to comprehensively model and understand that behaviour is a very important step in developing technology that is meaningful, positive and helpful. To do that, you need to be able to understand people.”
Oliver started in data science before big data was even a thing, with the idea that, “there is so much human behavioural data available to analyse and so much scope to make better decisions.” Back in 2004, she identified the mobile phone as the most personal of all computing devices. “I could see it was a technology that would enable us to understand people in a way we have never done before. We spend more time with our phone than anybody else in our lives,” she recalled.
The confluence of mobile’s surge in adoption with the development of a new academic discipline, computational social sciences, has bridged the commercial worlds of technology and academic social sciences which used to be hard to align. “Now we can have access to large-scale data about human behaviour and validate social science theories to identify factors that play a role in, for example, how cities behave,” she explained.
One of the areas that she has worked on for the last eight years is what she calls “Big Data for Social Good”, where she has explored four main areas using aggregated mobile network data that has been anonymised and made available for research purposes with a social good angle. The first is public health and tracking epidemics and pandemics. Said Oliver: “Because we can now measure human mobility, it is possible to model how infections spread. It is the first time in history that we can examine the mobility of the entire population. Even ten years ago, we couldn’t do that.”
Her second focus has been on emergency response to natural catastrophes like earthquakes and flooding, looking at how these affect local populations and how to optimise the recovery. A third area is economic development and inferring the impact of socio-economic factors, while the fourth is urban planning. “There are a lot of projects using mobile data for transportation modelling, hot spot detection, pollution, crime prediction and environmental planning,” she noted.
Mobile networks have provided cell traffic data, usually aggregated into days or weeks, within a number of projects and competitions. Orange’s Data for Development, Telefónica’s Datathon for Social Good, with the Open Data Institute, and Telecom Italia have all put vast data sets into the hands of academic researchers. Oliver said that, ”at the other end of the scale, there are small qualitative data sets, like MIT getting 100 people to allow their mobile phone and app usage to be tracked. But to infer for the entire population, you need scale and a representative data set.”
These research projects are typically multi-disciplinary, not least because of the need for complementary skills sets. “Data scientists are not experts in public health, so you need a team that includes epidemiologists, experts from the public sector, as well as the technical data people. It is very exciting, but also a challenge for academics used to working within a specific domain. The world is not siloed,” she said.
Bringing people together into a team is part of what makes for a successful project, especially given their different academic, ethical and skills bases. “It takes a while to set up a project,” admitted Oliver, “but when you do, the impact can be significant. It is a very exciting time to be in this area because the opportunity to have an impact is so big.”
When Oliver started out, the concept of using data science for social good did not exist. “What has happened since then is that is has been embraced by organisations like the United Nations, World Economic Forum, the World Bank - they see that the value in big data is to help achieve sustainable development and growth and to make better decisions.” But she added: “We are making progress, but we still need to educate people and show them case studies to understand how to realise that value.”