UCL Careers


Transition from academia to data science in civil service

Dr Zara Grout has a PhD in Experimental Particle Physics from the University of Sussex and now works as a Senior Data Scientist in the Planning Inspectorate (Civil Service).

Close up image of Zara Grout

1 June 2022

We asked Zara to tell us about her current role, her career journey, challenges of the transition from academia to industry and about any useful tips for researchers wanting to pursue a similar career. 

Tell us about your current role and organisation.

I currently work for the Planning Inspectorate – an agency within the Civil Service. Its parent department is the Department for Levelling Up, Communities and Housing. The Inspectorate has three main functions: Planning appeals, which people (or developers) can submit if they have had their planning application denied by their Local Planning Authority; National Infrastructure Applications, which are submitted directly to us for large projects on transport, waste, energy, etc., for example HS2 or Sizewell C; and Local Plan examinations, where we help local planning authorities ensure they have a plan for development, infrastructure, etc., in their area that is sound and complete. 

My role here is “Senior Data Scientist” as part of the Government Statistical Service. As the Inspectorate is very small, my work involves a wide range of activities including data and performance analysis, machine learning techniques such as natural language processing and clustering, and advising on related tasks, e.g., our data and software infrastructure and tooling, or working with the performance and reporting team. 

How did you move from academia to your current role?

After my PhD in Experimental Particle Physics, I did four years of postdoctoral research on the same experiment (ATLAS at the Large Hadron Collider) at UCL. During the last year, I became increasingly involved with UCL’s Centre for Doctoral Training (CDT) for Data Intensive Science (DIS) – this may have since changed name! I supervised a group project for 3 PhD students who were working with the ONS to analyse tweets and assess the topic and sentiment of them at scale. I also attended the machine learning lecture course to learn more about the theory and practical application of more advanced machine learning techniques. 

The CDT also ran a large number of events including student presentations, networking and seminars, which were a great way to familiarise myself with what was happening in academia and industry around this subject and talk to a wide range of people. 

As a result of one of these events, I arranged a three-month placement with The Guardian using data science techniques on their recipes data.  Shortly afterwards, I applied for my current job at the grade of Senior Executive Officer with the Planning Inspectorate, and UCL Careers were very helpful in my preparation for the application and interview!

What have the main challenges been in the transition from academia to industry?

I was already using data science skills daily whilst working as an academic, but without knowing the language generally used to describe them in industry. My involvement with the CDT, and the placement at the Guardian in particular, really helped me in being able to communicate my technical expertise. 

Working for a small government agency also meant that the emphasis of my day-to-day communication shifted from being very focussed on technical details to talking more about the ‘bigger picture’ or ‘why is this useful?’. Although I had experience with this from teaching and outreach, it involved quite a shift in how I presented my work and thought about which knowledge or insight could be most valuable to obtain from the outset. Overall, I think it has made me a better communicator and scientist, as I consider multiple perspectives early on in a piece of work and break down the findings to address them. 

Finally, compared to my experience with experiments, which are entirely focussed on collecting and disseminating data, organisations outside of academia often don’t think enough about what data they are collecting, how and where they are storing it, what quality it has, and what associated metadata is required, e.g., has this value changed in the last 6 months, if so, when and why? Could it change again? This was therefore quite a new challenge, but one which really benefited from understanding how these things could be done better and what difference it could make.

What does a normal working day look like for you?

I spend most time writing code for data processing and analysis using Python, usually in Jupyter notebooks, and usually producing graphs, tables and other visuals to share along the way. As with most data scientists, the majority of data analysis time is often spent on sourcing data, understanding the quality of it and any drawbacks, applying required cleaning and pre-processing and often merging different sourcing together to serve my purpose. Some of this is done collaboratively with colleagues in the data and performance team.

I also spend a significant amount of my time problem solving, teaching and supporting others who are less experienced with coding or analysis. I also spend plenty of time learning things myself, either by taking courses, reading and learning independently or attending conferences or presentations. 

I often have meetings with colleagues in jobs across the inspectorate, such as planning inspectors, support staff in HR, customer services, digital services or the Innovation team to see where we could work together or advise them on data analysis. I recently worked with the Innovation team on a discovery project to see how AI could help inspectors deal with the large volumes of written documents they receive and review for Local Plan examinations. This involved selecting and working with a small team from a data science consultancy to develop prototypes, test out lots of ideas and assess how effective they could be. 

What are the best things about working in your role?

There is a lot of variety in terms of the people I work and speak with, the data I deal with and analysis techniques I use. Projects tend to take place on shorter timescales than an analysis for an academic paper (or certainly those I worked on previously on ATLAS, which were 1-2 years). Also, I receive more feedback on whether my work is understandable, helpful and interesting, as I tend to share it more frequently (partly to address specific questions). 

At least in my experience so far, personal development and soft skills are recognised as very important to conducting my more technical work effectively, and it is expected that I should have time to develop these at work, rather than feeling I need to squeeze them into my personal time, as was sometimes my experience in academia. 

There is a lot of flexibility in terms of how I want my role to look, and opportunities for me to focus on aspects that I find more interesting or want to develop further.

What are the biggest challenges?

In such a small team, I sometimes miss how easy it was to discuss complex analysis problems and results with colleagues, even if they were working on something different to me. In particular, I have had less technical ‘supervision’ from my line manager and have had to seek advice and feedback from the wider government data science community when I encounter specific problems or need a second pair of eyes. 

Data analysis is also fairly new to the Planning Inspectorate, so one of the biggest challenges has been advocating for software and tools required and to help improve data literacy and people’s mindset about data. I’ve often enjoyed this, writing blogs, presenting at conferences, etc., but it can sometimes be frustrating when things take longer than they need to, or decision makers don’t recognise the resources needed to take advantage of the data team.

Is a PhD essential for your role? If not, did you find your PhD experience nevertheless useful?

No, there are definitely other routes to my job, either through industry or within the Civil Service. However, a PhD enables you to enter at a higher starting grade. Also, experience from my PhD and postdoctoral positions has proved invaluable, both for technical skills and higher-level critical thinking and softer skills, such as organisation, proactive learning style and working with others. I only appreciated some of these skills fully having encountered people who did not have the same experience as me in the workplace!

What’s the career progression like for your type of role?

There are lots of options for a Senior Executive Officer (SEO) data scientist in the civil service. Data science is up and coming in government and something I particularly like about this path is the ability to progress whilst retaining a lot of technical activity. I am currently in the process of transferring to a Grade 7 (one grade above SEO) data scientist role at the Cabinet Office, which still involves the activities I carry out now but at a higher level. There are also Lead Data Scientists above that. 

Generally there is more line management, etc. involved in the higher grades, but people still do their own analysis and research, if they wish. There are also options to go for less technical progression and focus instead on e.g., management or strategy of data teams, departments or more widely across government, or to shift sideways and consider data architecture, data engineering, software development, etc. All of this is also very transferrable to private companies or consultancies.  

What top tips would you pass on to a researcher interested in this type of work/career?

There are a lot of options in data science, both in the civil service and private sector, and I found speaking to friends, friends of friends and people at networking events really helpful in deciding what kind of role and what kind of organisation were most interesting to me. It also gives you a chance to ask questions and get some exposure to the language people naturally use in these roles! 

For the civil service specifically, applications can be quite involved, which generally makes the process fairer in my experience, but you should allow plenty of time for thinking through how you address the job specification and required behaviours, skills or competencies. 

UCL Careers were fantastic at helping with this, including practicing the “STAR” method which didn’t come naturally to me! In terms of technical skills, Python is very transferable but you will find other languages used in various roles – often your interest and ability in learning programming languages quickly and applying them to problems is more important than having used a particular language or method before.