news    views    podcast    learn    |    about    contribute     republish    

Juan Mateos-Garcia


Homepage | Github

I am the Head of Innovation Mapping in Policy and Research at nesta.org. My job is to use new data sources and analytical methods to improve innovation policy and practice. I am particularly interested in how new technologies and industries emerge, on the way in which ideas spread across networks, and on the processes through which, as a society, we can manage this process of continuous change for the benefit of all. Technically, I am interested in the potential of machine learning and network science as tools to understand our complex economy, and of reproducibility as a way of building trust around new data sources and methods, making them more suitable for policy application. I use Python and R. I am currently leading Arloesiadur, a project to build a data analytics platform to inform innovation policy in Wales. Other examples of data analytics projects I've worked in include TechNation 2016, several Nesta data blogs using new web datasets such as Meetup, GitHub or Kickstarter, a project to Map the UK Games Industry using big data, and the Net Effect, where we used Twitter data and social network analysis to measure connectivity at innovation events Previously, I worked on a programme of research and policy development to understand data science skills, Next Gen, an independent review of education and skills for the video games and visual effects industries, A Manifesto for the Creative Economy, which proposed a plan to help the UK creative economy remain a global leader in digitised creative markets, and Creative Clusters and Innovation, which created the first geography of the British creative industries.



Dig below the surface of some of today’s biggest tech controversies and you are likely to find an algorithm misfiring. These errors are not primarily caused by problems in the data that can make algorithms discriminatory, or their inability to improvise creatively. No, they stem from something more fundamental: the fact that algorithms, even when they are generating routine predictions based on non-biased data, will make errors. To err is algorithm.