In our new blog series, Industry Leaders on the Future of Data Science, we asked industry leaders from across the UK what they expect to see from data science in 2022.
In this edition, Christina Boididou, Data Scientist at the BBC, talks about the future of AI and human teams and how they can be combined tactfully and ethically. Christina’s work involves developing machine learning solutions for BBC products, focusing on the recommendation systems, communicating her work across the business, and ensuring that it aligns with the BBC’s editorial values.
Christina, thanks for joining us. You’re a data scientist at the BBC, from your perspective, what should data scientists and businesses expect from 2022?
Data practitioners are part of an extremely interesting industry, yet very fast paced and complicated. One of the biggest challenges in data science is the data itself. Whatever people resources and variety of methods might be in place, the data quality is the first and foremost for successful data solutions.
Data scientists need to ensure that they have good quality data before they invest in complex implementations; knowing the data and its limitations should be the priority. To overcome data challenges, they should communicate with the business side and highlight how correlated is the quality of any data solution with the data provided.
Apart from the data, data scientists must deal with the number of specialisations that are coming up in the industry. Data science, as a relatively new and at the same time vast field, doesn’t have clear boundaries between different roles within it. Different specialisations arise and practitioners entering the field have the challenging task to find their own place; it is important to remember though that overthinking about picking a specialisation and sticking to tools and terminology is not always a good way forward. Data scientists should stay flexible and think about their own vision rather than sticking to typecast themselves.
How can people prepare themselves for these changes and challenges?
Everyone who is involved in data science projects should stay adaptable to what the future brings. Both individuals and organisations should see things with a critical eye and cultivate their ability to look at the bigger picture of a problem. Although specialisation is the focus right now, this can’t solve the larger problem effectively most of the time. People should prepare to stay agile and make sure they communicate between the different components that participate in a data science challenge.
We’d be remiss not to talk about the challenges of managing remote teams, with the data science industry perhaps benefiting the most from work from anywhere. What’s your advice for managing remote teams?
Due to the pandemic, the working normal has shifted to remote working. This created opportunities for both companies and individuals. Companies opened up their candidate pool and data practitioners have now more offers which are not limited to their local area. In the longer term, it is highly likely that this is going to be the new normal; working lifestyle, compensations, relationships between colleagues, even socialising and living will change based on the new situation. The industry will adapt to this new reality but at the same time affect it and consolidate it.
By default, managing people is a very challenging task, let alone managing in a remote setting. There are various points that need attention by the managers in that case, some of which are organising everyday tasks, having conversations about peoples’ personal career objectives, thinking about each person’s personality. For the former, I think it is crucial to put processes together to ensure there is a smooth collaboration where everyone feels happy and included. Work and tasks should be communicated with a preferred way to make sure that everyone is up to speed.
It is a tricky one, because managers should both give space to the team and trust them but at the same time show they support them in whatever they are trying to achieve. For the career objectives, it is also important not to neglect them in an online environment. Being in the office facilitated conversations and networking, but this became harder with remote working. Managers should be much more engaged with the team members to make sure they are still aligned with their future path and encourage them to progress. What should not be forgotten in this situation is the personality of people. For example, managers should think how they work with introvert and extrovert people and accommodate their needs.
In any case, there is not a unique recipe that exists to be successful at this; everyone should be open to experiment with different models and processes.
What are your tips for new data scientists joining remote teams and companies?
Starting a new job fully remotely especially if it is your first job can be a bit challenging. It is harder to familiarise yourself with a company’s practices when working remotely because you miss the everyday informal chats and the atmosphere in a physical office. However, I believe that in those cases the key point is to be proactive; I’d suggest new data scientists to communicate constantly with their colleagues, engaging in conversations about their work and the next steps in a project.
Asking questions and sharing results early can also be a catalyst to a good remote working collaboration, minimising mistakes that can arise from absence of communication. Also, it is essential to be deliberate about catching up with your colleagues or wider team, even if it feels unnecessary in the beginning. As the office doesn’t exist with the format we knew, we need to recreate conversations the way we had them before. At the end of the day though, it’s a new situation and all of us are learning along the way.
Diversity and Inclusion are, rightfully, taking their place as part of the wider conversation around the industry. How can we make data science more diverse and inclusive?
When businesses think about diversity and inclusion, they often think about it as a should-have and not a must-have. However, businesses whose products are based on data science need to realise how thoughtful they should be about their strategy, not only because they should give opportunities and consider diverse groups of people, but also because of the benefit this could bring to them. When building a new product or service that speaks to a wide audience, having employed people with different skills and from different backgrounds can only lead to a more complete product. Especially when it comes to a data product, this need is even more emerging because data is complex; it can be interpreted and used in different ways.