The last decade has seen data science make huge leaps for both people and planet. Not least because, thanks to significant developments in how datasets are captured, stored and processed, we now have more data, in a more usable form, than ever before. As Unai celebrates its tenth birthday, I thought I’d take the opportunity to look back at the last decade of developments and ask, What has data science ever done for us?
1. Fighting pandemics
What could be more topical than the role of data science in fighting pandemics? It is, without doubt, one of the biggest challenges to global health and national economies for a generation, and data science has played many roles. Everything from finding a treatment plan for COVID-19 to searching for a safe and effective vaccine involves the storage, transfer and processing of huge datasets. Behind the scenes the UK’s NHS is heavily reliant on its data science teams to help understand where infections are happening, which treatment pathways are working, what demand for their services will be and how they need to schedule staff and manage logistics to successfully combat the pandemic. It’s a mammoth effort, underpinned by fast analysis of huge datasets.
2. Genomics
In 2013 Genomics England was founded to deliver the ground-breaking 100,000 Genomes Project. This flagship project had four key aims: to create an ethical and transparent programme based on consent, to bring benefit to patients and set up a genomic medicine service for the NHS, to enable new scientific discovery and medical insights, and to kick start the development of a UK genomics industry. The project overcame extraordinary data challenges and paved the way for mainstream genomics medicine services in the UK.
3. Financial services
With the UK’s open banking regulation coming to into force and technology developments supporting the unbundling of banking services, the path has been smoothed for so called ‘challenger’ fintech companies to spring up and offer specialised services. Data science has supported advancements from chatbots to decision engines which help to address the needs of smaller market segments, particularly those who are aren’t properly served by larger financial institutions. Of course, this is a contentious issue, and certainly one where ethics and governance has an increasingly large part to play.
4. Transport
Data science has bought the ability to process millions of images at speed, without which the autonomous vehicles starting to join our roads would be unable function. The jury is still out on whether autonomy really brings safer transport, but early studies (such as one published by Tesla in 2019) suggests that they certainly could, finding that autonomous pilots only had accidents once every 3 million miles, verses once every half a million miles for human drivers.
5. Energy use
Sadly, the last ten years have seen a rapid increase in global warming, the melting of the ice caps and CO2 emissions. The story isn’t all bad news though, and data science has certainly played a part in what I hope will become a global change in how we use, and produce, energy. The UK, for example, has decreases its carbon emissions by as much as 29% (according to Carbon Brief’s analysis). While this is largely down to a change in the type of fuels being used, a significant proportion is also due to data supporting previously-unimaginable efficiencies in the production, storage and use of power. One interesting example is data-enabled software platform Kaluza, which is part of Ovo Group. The platform uses machine learning and AI to create a more flexible energy system, optimising devices to use energy off-peak, when costs and carbon levels are lower. These advances are particularly important as energy production is increasingly from renewable, decentralised sources.
6. Digital Healthcare
The growth of devices and wearables monitoring everything from our step count to our heartbeat, serious health conditions can now be predicted before they become irreparable. One example I am particularly excited about is OKKO Health, who use apps to gamify the process of measuring vision, generating a longitudinal record of users’ eye health for clinical use. Data science has a huge part to play in the future of these health tech companies as huge volumes of data help detect not only health deterioration but also previously-unrecognisable patterns that can predict decline at the earliest possible moment – when intervention is still possible.
7. Diversity
There is intense scrutiny and debate over the extent to which productionised data science is modelling inherent bias within datasets into systemicly-biased decision making process.
However, the last decade has also seen data science used to tackle social inequality head on by identifying specific features of these inequalities and informing interventions to help close them. Several organisations have sprung up to address quantifiable inequalities, such as GapSquare who use data science to process hundreds of thousands of payroll datasets, revealing pay gaps between different genders and ethnicities and identifying ways in which the gap can be closed and real pay equality can be found. I hope the next decade will see more organisations such as data2x, which works through partnerships with UN agencies, governments, civil society, academics, and the private sector, to identify areas where massive swathes of data on women and girls is missing or incomplete in the algorithms that make decisions about how we live and work.
8. Social mobility
Social mobility has been reported as being stagnant in the UK for the past 5 years. Data science is helping develop targeted approaches to improve social mobility, with organisations such as MyBnk, a non-profit who aim to deliver financial education and improve financial literacy in young people in the UK, using data science to better target their approach. There is evidence that data science can be used to aid urban planning policy interventions and identify areas and communities where a difference can be made. Whilst there is a lot of work still be done in this area, the ability to understand the nuances of this problem is improving and hopefully paving the way for real progress.
9. Cultural Heritage
When it comes to attracting visitors, why should online retailers get all the good data science? Despite the different sets of challenges, organisations offering real world interactions are still successfully delivering a 21st century experience and 21st century insights. The University of Exeter’s Vista AR project, for example, is working with cultural heritage attractions across England and France to trial portable devices loaded with Augmented Reality content to bring the site to life. The reviews and digital traces generated by people using these devices as they move through the sites are being analysed to help improve the experience and focus investment.
10. Education
The developments in artificial intelligence and data science over the last decade have also enabled us to approach a point where we can use it to better understand our own human intelligence. The last decade saw a computer beat the world champion of the complex strategy game, Go (a popular pastime for many of the Unai team). But AI has the potential to not only help us asses and measure our intellects by the traditional yard sticks of applied knowledge, but to also help us understand our emotional intelligence better, from how well we work with other people to how resilient, self-aware, motivated and self-effective we are: themes explored by Professor Rose Luckin co-founder of the Institute for Ethical AI in Education.
So, next time you wonder, “what has data science ever done for us?” these are some areas to explore. Making big decisions with big data is a big responsibility, and there is certainly a need to ensure that such responsibilities are handled with ethics at the heart. Despite the challenges, I am hugely optimistic about the opportunities that data science can bring today’s pioneers and the benefits it will bring to people, and planet, over the next decade.