The judges selected eight ideas for further development by our project teams, ranging from data dashboards to service automation via Big Data and AI applications.
The teams are currently researching these ideas in advance of the ‘Dragon’s Den’-style shortlisting event in October: below we present a brief synopsis of each one. You can also view all the ideas we received on our All Ideas page.
Under the government’s ‘Levelling Up’ agenda, some 22,000 Civil Service jobs are expected to move out of London and the South-East over the coming years. An online tool showing the current and planned geographical locations of Civil Service teams and specialist staff would help Civil Service bodies to plan their relocations, forming ‘hubs’ in particular fields to realise economies of scale and develop professional communities.
Idea submitted by Ben, BEIS
Some recipients of the Personal Independence Payment are entitled to a 50% VED (road tax) reduction, but the process is entirely manual and quite complex: claimants must show their PIP receipts along with DVLA documents and an MOT certificate, and the process takes 4-8 weeks. Systems for initiating and modifying VED discounts could be automated by linking DVLA and DWP systems, where all the required information is held. A linked idea proposes asking PIP claimants for data on their medication regime, passing this information to DVLA to hasten and improve decision-making on driver licensing.
Idea submitted by Sarah, DWP and Sean, DVLA
By linking together existing datasets, departments could reduce their losses to fraud and error – spotting when benefits claimants have been out of the country for long enough to lose their entitlement, for example, and identifying individuals who are under-reporting their income. These systems include DWP’s Customer Information System, HMRC’s Construction Industry Scheme and the Home Office’s API border management system.
Idea submitted by Aaron, DWP
Create a dashboard showing estimates of the numbers of UK citizens living, working and holidaying in cities and countries around the world – supporting consular service development, emergency planning and disaster response work. Data sources could include social media, transport and travel providers, and perhaps telecoms operators, as well as public bodies.
Idea submitted by Tom, FCDO
The UK’s peatlands are currently drying out, emitting around 20 million tonnes of CO2 annually. Blocking moorland drains can reverse this process – but locating them on the ground is a difficult and labour-intensive process. Deploying innovative ‘Generative Adversarial Networks’, we could rapidly identify drains using aerial photography, much reducing the cost of protecting these important environments.
Idea submitted by Damien, Defra
The UK’s medical supplies reach patients through a complex supply chain, creating risks at times of disrupted travel and/or spikes in demand. By bringing together data from public bodies, transport and logistics firms, and medical and pharmaceutical suppliers, we could build a clear, up-to-date picture of our current stocks and supply lines – supporting action to plug any emerging gaps before frontline services are affected.
Idea submitted by Helen and Graham, DHSC
We know that prisoners who maintain close family ties during their sentences are less likely to reoffend, but data on their social contacts is currently dispersed and of variable quality. Bringing this data together and improving its collection would enable prisons to plan additional support for prisoners suffering from a lack of social contact, helping to reduce reoffending rates.
Idea submitted by Charlotte, Ministry of Justice
A synthetic data approach uses AI to learn the ‘shape’ of real datasets and generate new artificial data with similar characteristics: in computer imagery, for example, AI can use pictures of real people to generate new, realistic human faces. Building on the work already carried out with synthetic data in DWP, this idea proposes the creation of cross-government synthetic data repositories to allow ideas and policy proposals to be tested without using real personal data.
Idea submitted by Shruti, DWP