Dan Fitzgerald and Connie Becker, Analysts within ECC’s Data and Analytics team, reflect on their recent experiences developing new skills…
Were you one of those kids that spent hours playing The Sims in the past? If so, this is definitely the blog for you! And if not, keep reading, we may convince you that Sims is useful in developing your analytical skills!
A few years ago, we would never have thought that all those hours building houses, deciding what our Sims will need to fulfil their lives, or managing virtual families, would’ve been foundational to one of the skills we needed in our roles as Data Analysts.
So, when we decided to apply to different learning and development programmes - Dan for the Health Service Modelling Associates Programme and Connie for the ONS Data Science Accelerator Programme – we were surprised that simulation was going to be a skill that we would end up developing.
As you can probably guess by now, something about these learning programmes felt oddly familiar. Just like in our favourite game as children in the noughties, we had in our hands a complex system (Adults Social Care and Health) that we felt could be better understood by doing something a little different.
However, reality brings far more complexity than our games ever did. If you’ve ever interacted with the health and care system, you’ll know it’s very complicated. Complexity can be hard to translate into an analytics model, even more so when you’re learning a new method such as simulation. Understanding how adults move within the social care system was the first challenge. Understanding how they interact and transition between health and care was another entirely. How would we go about building a simulation when we had far more questions than answers. This is where our respective programmes helped us.
Dan started his programme delivered by the PenARC (a research body comprised of the NHS, local authorities, universities, and others). That was six months ago now, in that period he was able to attend weekly lectures provided by experts in different areas such as NLP, simulation modelling, system dynamics – the list goes on and on. These weekly sessions involved lectures in specific methods and how to employ them in Python with examples applicable to health and care. He was encouraged to work collaboratively with peers from other organisations to find novel solutions to common problems.
The course was of particular interest due to the roll out of health and care data platforms by Integrated Care Boards (bodies tasked with planning and funding local NHS services) across the country. These platforms allow us to use data from health and social care to understand questions like which cohorts of patients are using services and are these services meeting their needs. We hope now to use the methods from the course to understand entries into adult social care based on different health conditions. If we know that cohorts with certain conditions or certain combinations of health conditions enter adult social care not long after diagnosis, we can perhaps then start saying “well, how can we act to enable independence for longer?”. HSMA methods will be vital for that. Having this structured way of learning within the course meant that these new techniques that had first appeared to be entirely alien were broken down into easily digestible chunks that can now be employed in our everyday work.
When Connie began her programme delivered by the ONS, she expected that she’d employ a traditional machine learning approach to tackling her business problem to develop tailored care plans for adults in social care, supporting them to live as independently and safely as possible, enhancing their quality of life while ensuring their wellbeing. By creating a simulation framework however, we can better understand what sort of services an adult could potentially need in the future, an important tool to estimate what resources are allocated to best support individuals.
The benefit of the programme Connie took part in was working closely with expert mentors from a wide range of backgrounds across the country. That meant that Connie was able to draw on their expertise, and work with her mentor to employ a new method that was better suited to answering her business problem, agent-based simulation! By the end of the programme, Connie was able to create a simulation framework from scratch, coding everything without fancy R libraries! It was hard work but very rewarding and built confidence in newly learned simulation modelling skills.
Another benefit of the programmes was networking. Colleagues in other government departments, the NHS, and local authorities were just as intrigued to learn how the simulation methods are a good alternative to traditional machine learning when it comes to decision-making. Building this data science community across the country means we can pool our expertise and improve all our services to the benefit of our respective service users.
Simulation has brought us together. As kids, we both enjoyed spending countless hours in the world of the Sims – building houses and determining futures. As service providers in the real world, we must anticipate and respond to user needs and external pressures to deliver our service. Simulation can allow us to understand how users interact with those services, perhaps in ways we hadn’t expected. The only way, we’re able to adopt these novel methods is to continually hone our craft as data professionals and we are fortunate that we are in a team that values innovative development of skillsets in a way that is supportive of these emerging techniques.
Leave a comment