In conversation with Emma Presley
Emma Presley is the Lead Data Scientist at the supporting Service Design for Universal Credit at the Department for Work and Pensions. Emma will be joining the panel discussion Maximising the value of data science to enhance public service delivery at 11:50 on Tuesday 25th June 2019.
What is the biggest challenge in your role at DWP?
My team and I exist to ensure that data generated from Universal Credit’s digital systems is used to inform how to develop the service. Working on a large scale digital programme that is constantly improving means that our working context is constantly shifting, and there’s a high barrier to understanding the data generated from such a complex system. As data scientists, we are embedded within the development teams that we support; this close multi-disciplinary working helps us keep up with the pace of development (just about!). It’s a fast-paced and dynamic environment to work in, but the context of the problems we are working on makes it extremely rewarding.
What are some of the successes in the use of data at DWP that you are part of?
I am very proud of the data-driven culture my team has managed to foster across the Universal Credit digital programme. We have gone from being a small centralised group delivering ad hoc insights, to one where data scientists are embedded within product development teams and delivering data products that support and inform the delivery of a live digital service. In a recent internal showcase, it was heartening to see non-technical colleagues consistently advocate for the use of data in their decision making, with multiple techniques including text analytics, clustering and semi-supervised machine learning woven into the narrative with ease. This is not just testament to my team’s ability to tackle these kind of techniques, but also their ability to bring them to life in a relevant business context (which is easy for me to say!).
Why do you think so few public sector organisations reach their data science potential?
With the digitisation of public services comes a number of new challenges for public sector organisations to overcome. Digital services are typically powered by data structures designed to optimise the system’s operational efficiency, and often produce large volumes of data at high velocity. Whilst this type of data is a hugely rich source of information and insight, it is not designed to benefit your typical analytical user! A lot of public sector organisations are used to dealing with well structured, low velocity clerical data presented in an enterprise system; wrestling value from an unwieldy, constantly shifting behemoth of digital data sat up in the cloud is a distinctly different challenge! As a result, lots of organisations are facing a sizeable skills gap, which I think should be addressed by investing in the development of the extremely knowledgeable and highly passionate employees who are likely already working for them.
Where would you like to see innovations in data science at DWP go next?
I would love to see examples of applied data science being embedded within the products and services that DWP provides. An example of this could be the triaging of responses to online messages from citizens, so that our operational colleagues can prioritise their time to where support is most urgently required; we get a lot of thank you messages, which are lovely to read, but could be deprioritised for response against other more time-sensitive messages. This type of approach relies on multi-disciplinary working, an appetite to build and fail fast in an Agile working environment, having the right tools and technologies in place, and most importantly keeping in mind who you are building for in the first place (spoiler: the citizens who are your end users!).