00:00 Ladies and gentlemen, good day and welcome to FNA Talks: A Technology Update with FNA’ers and friends. My name is Adam CsabayI am the Suptech Lead at FNA and it will be my pleasure to guide you through this episode. At FNA Talks, we are drawing on the expertise and experience of key Fintech, Regtech and Suptech authorities to discuss the trends and developments defining the technology and the innovation landscapes. 


In today’s special episode, we will focus on the recently published book, entitled, Data Science in Economics and Finance for decision makers and I am very pleased to say that we are joined by one of its co-authors, Adrian Waddy from the Bank of England. 


Adrian, welcome and thank you for joining us at the FNA Talks. 


Thank you, Adam, and thank you for including me. 


1:10 It is our pleasure Adrian. Adrian, to kick off our discussion, could you please briefly introduce yourself and tell our listeners about your background? 


Of course, so I currently work for the Bank of England. I was based in the technology department, although at the moment I’m seconded in the PRA, which is the regulatory arm of the Bank of England, into the Data and Innovation team. But previously, while working in technology I was the technical lead on the design phase of the implementation of a Data platform for the Bank of England and that was broadly the subject matter of the chapter that I contributed toward the book. 


1:50: Thank you Adrian, now turning towards the focus of your contribution to the book. What does your chapter cover and why would you suggest that this topic is important?


Yeah, sure… It covers in broad, non-technical terms what’s involved in deciding whether to, and some of the crucial decisions that need to be made about implementing a big data platform. We help with a guide to sort of identify whether the data you have is big, and we also help with a discussion about whether you implement in the cloud or on your own premises or maybe a hybrid of the two. And then we talk a little bit about standardizing approaches to ingesting, processing data can yield benefits and touch upon the options you have for presenting data and the different types of personas that are interested in consuming data. 


3.00: What would you argue are the two main insights for demystifying Data Science? 


Well, I think the crucial question is, you know implementing a data platform is not necessarily a small undertaking and identifying if that is necessary to deliver what is required. Increasingly that is the case in central banking, in regulation, in financial services, but it's not always the case. It might be that your data doesn’t fit the definition of big in terms of having volume or velocity that is required for a large platform or indeed the variety. However, if you do think that it looks fairly convincingly like you have a need for a big data platform, there are a number of questions then there are a number of questions that you need to follow. One of those, that is not necessarily related to big data, is whether to start your implementation in the cloud or on premises within your organisation. I think that question is true regardless of any large technology project. However, because it's so crucial for big data and we do spend a small amount of the chapter talking about the options that are there, I think that’s crucial. And we touch upon in some detail about the opportunities that exist for standardizing approaches and therefore being able to implement new systems, implement models, implement new analytical outimes in a standardized approach which allows us to do it far quicker than we might do if we were building from scratch each time. 


4.42: Many thanks Adriand for these very interesting insights. Let us now try to summarise our listeners some of the key takeaways. What are your two pieces of advice for decision makers starting Data Science? 


Yeah, yeah that’s a really good question. So, I think the two things here are, this is a very fast moving environment. Like any environment you have a roadmap and your plan, but don’t be surprised if the details of your plan have altered even two years down the line. Because an awful lot of the big data implementations are based on open-source tooling and that environment can change very quickly, tools and development’s develop very quickly. So what looks like a certainty for you to be using in two years might have been superseded by the time you get there. So, do plan but be flexible. Be prepared to learn both, you know in a traditional, academic, didactic fashion but also by partnering with key organisations. So that would be one key thing. And the other is, although on the face of it, the implementation of a big data platform is a technology project, I would say of at least as much importance, would be your target operating model that surrounds that platform. So here, I’m talking about the people, and processes and the governance that you wrap around the platform to ensure that it is used efficiently, to ensure that you understand the data that exists in there and to ensure that people engage with using the platform. So they’re are my two insights, Adam, 


6:37: Excellent, thank you very much Adrian. And what are the two main lessons you would draw from your chapter? 


Having gone through this with the Bank of England, I think one of the key things is that it is difficult. It is hard, so perseverance, tenacity is very, very important and being, you know traditionally, if you used a Microsoft BI platform then, you know broadly speaking, there wasn’t a lot of difference between 2005 and 2015. It was broadly sql server and analysis services and you know perhaps some Power BI later on, but that same, in a big data world, the tooling world changes very quickly, so it goes back to that two year horizon, so don’t be surprised if things change. I would reiterate the importance of the cultural change, the people and process change that needs to go with the technology implementation. 


7:49: Thank you very much Adrian. Before we conclude our discussion, let us still spend some time focusing on the future. Could you please share with us what is your vision for Data Science and its evolution? 


Yeah so this is really interesting particularly as I work now in a Data Science community and I think it's sort of, the common term is democratization of Data Science. I think we are getting an increasingly savvy user-base, so there still exist users who just want to press a button and have a canned report delivered to them and we mustn't lose sight of that. However, we’re getting increasingly data-savvy users who understand a bit more about the complexities of delivering the data but also want to bring some of the tools. So more and more often, I’m hearing a request for people to be able to access and process large amounts of data with Python or using Spark. And I think the future of DS is going to be how a greater proportion of our user-base within regulators will be DS-Savvy and will want to bring DS skills, tooling and thinking to the data that informs their decision making. 


9.16: Many thanks for your very interesting responses Adrian. You have provided us with some great insights and I’m very much looking forward to reading the book as well as your chapter. 


Thank you Adam 


Ladies and gentlemen, many thanks to you for your attention. If you have any questions or comments for Adrian and I, please let us know on Social Media or write to us at Adam@fna.fi. I very much look forward to reconnecting soon for another FNA talk and in the meantime, have a good day, stay safe and goodbye.