This Autumn, in collaboration with ASI Data Science, Future Cities Catapult launched a series of fellowships that applied data science skills to defined problems in UK city regions. We believe that the use of applied data-driven research in city regions will develop the role of cities as intelligent customers and enable them to engage the market with confidence when investing in innovation. It’s now November and our first cohort of data science fellows have finished their 6-week placements, which yielded insights into two different city challenges.
In Newcastle, our fellow, Dr. Raquel Vaz, worked on a data challenge around air quality. Annual averages for NO2 in this region, along with many UK city regions have exceeded EU limits for the last four years. Raquel worked with the City Council and the Urban Observatory to determine what the impact would be if the council restricted access to vehicles on a key city centre link, Blackett Street. Raquel used time series analysis and geospatial statistics to demonstrate that temporary road closures significantly improve air quality in the centre of Newcastle during selected weekends.
Raquel’s research also unearthed a roadmap for potential future analysis and future data sharing, which could lead to systematic improvements in air quality in the long term. Her research highlighted that to assess the full economic impact of such a policy, data scientists would need access to further sources of data and certain other datasets held by the council and others. Examples of this data include people movement data, vehicle counts and vehicle journey times. Access to and sharing of data remains a wider issue across local councils, service providers and cities and this must be addressed if they are to truly realise the benefits of data-driven services.
In Nottingham, Dr. David Hopkinson worked with the council’s GIS Team to look at the area of private rental and the council’s landlord licensing scheme. His research has helped Nottingham City Council to enforce a new licensing scheme that ensures privately rented properties are in a good condition for tenants. Using a machine learning approach known as ‘Active Learning’, David and the team in Nottingham were able to combine data science and expert domain knowledge to make the most of the limited data they had available and identify properties that were unlikely to be registered with the scheme but needed to be. The information generated by this approach helped to more efficiently identify homes that may not have signed up to the landlord licensing scheme and were considered most vulnerable. Early estimates suggest that this active learning approach could create a six-fold efficiency increase in identifying appropriate rental properties for the council’s Safer Housing team over the next two years.
In both city regions, work for the fellowships has been well received and the learnings and methodologies created by the fellows could be transferred to different regions with similar challenges. As a Catapult we are now looking to refine these learnings and make them available so that other city regions can benefit from the insights of our fellows.
As we mentioned in our first fellowship blog, we’re currently looking into how we can expand and develop these fellowships and help different city regions benefit from applied data-driven research. If your city is interested, then get in touch with Yalena Coleman, our Head of Accounts at Data & Demonstrators. Alternatively, if you’re an organisation looking to unlock city data to support your application or service, let us know and we can explore an opportunity for you, our fellows and cities to collaborate.