https://openpolicy.blog.gov.uk/2014/09/18/data-science-2/

The Data Science in Government programme: progress so far

August marks the mid-way point for the data science in government programme. We posted our first blog about our work in May and we would like to use this opportunity to update you on the progress we have made up to now and the next steps until autumn.

Over the last 6 months, we have completed the first phase of our alpha projects which show some really interesting findings and a basis for future work. We will continue to work on some more in-depth projects as well as further short alpha projects. We are also stepping up our engagement inside and outside government to discuss and develop plans for data science in government; the kinds of projects where data science could really help to improve the way we develop policies and services and what this means for us about our skills, technology and other practical needs (including ethical considerations).

Alpha projects

The alpha projects we have completed so far demonstrate the range of applications across government and the different approaches that data science offers us. Some projects take advantage of strong visualisations using open source tools and data science techniques to help the user to see new patterns and insights. Other projects have taken this visualisation aspect further to create an interactive tool which can be interrogated to help understand where service improvements can be made to affect cost and quality. Other approaches make use of sophisticated algorithms to gain deeper insight to help rethink services. Several of these projects are available to explore online and we will continue to post about completed projects here.

One particular project involved working with Foreign and Commonwealth Office (FCO) to help ambassadors understand their physical and digital networks and discover distinct communities in contact circles. Network maps visualise connections of twitter followers and the groupings created can show political affiliation. These maps have given FCO insights to help focus their online engagement and they could also help policy makers understand social media landscape in real-time to improve engagement. There is potential to reuse this in other scenarios across government as well as to think about other potential insights. For example, it could be developed to look at tweets and retweets or identify under-represented groups for instance, in consultations.

Intricate map showing connections between twitter feeds and searches around FCO themes
FCO alpha project network map (click to zoom)

Capability

Our second strand of work is to identify and build upon data science capability in government. Through our engagement with teams, departments and a network of practitioners, we have discussed how we can best develop and encourage data scientists and analysts.

Three main areas have so far been flagged as needing attention: technology, time and training. Current technology used within some government departments can present a challenge to using data science techniques and to get the full benefit, we need to find new ways of working. Analysts also need time to train and learn these new techniques as well as the time to explore their data and experiment with different ways of getting insight from it.

An outcome of these discussions has been to initiate an ‘accelerator’ programme; this will provide a small group of analysts from departments with the technology, time and training they need to work on a short alpha project. They will receive help from data scientists in GDS and time away from their day jobs to work on new, innovative projects. The outcomes of these projects will be reported towards the autumn and we hope to use the findings to help us with future plans for how we develop the skills of our analysts.

Unlock practical barriers

Our practitioner group, our conversations across government and our alpha projects have provided feedback which highlights areas that could create difficulties for a data science project.

For example, data science gives the ability to layer data sets to see insights that would not previously have been found in one single data set. However, the sharing and linking of data in this way raises some policy and ethical questions which should be addressed. We are working closely with the data sharing policy team in Cabinet Office who are currently running an open process to develop data sharing policy in collaboration with interested parties outside of government.

Next steps

Over the next few months we will continue our pipeline of alpha projects, with the addition of the analysts in the accelerator programme who will sit alongside our data scientists in GDS. We intend to increase our engagement both internally and externally to help develop our thinking on the practical barriers and particularly the ethical issues raised by data science techniques. As the projects and ideas develop, we will continue to blog about our progress and provide access to the tools and analysis we develop. If you would like to find out more, please contact us.

Leave a comment

We only ask for your email address so we know you're a real person