Demystifying Data Science, Part 1

One of my aims is to continually improve finance engagement with new tech by demystifying technology and moving finance professionals from fear of the unknown, FOFU to fear of missing out, FOMO!

Below I start on Data Science (not a deep techie explanation) and give short definitions against a finance and accounting example.
Firstly, what is Data Science? In a nutshell!

Well, there isn’t a standard definition of Data Science, but its the sexiest job in the world according to Harvard, and “a multi-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from data in various forms, both structured and unstructured, similar to data mining.”, according to Wikipedia.

Finance and accounting example: Mining customer payment data and external credit data to create a software engine that can predict how likely a customer will pay and when, and/or the associated bad debt provision you should provide.

How do we “do” Data Science?

Normally run as multi-disciplinary projects looking at specific business issues/problems (note: we start with the problems, not the tech!).  By their very nature, projects are highly collaborative with a diverse expert team, including the following disciplines and suggested roles.

• Business expertise – the accountant or finance business partner
• Maths, statistics and probability – the data science dude
• Software programming – the data developer
• Communication – the data analyst or finance business partner
• Note: no data scientist or accountant is an island!

Oh, does this mean there is an existing role for me in Data Science?  

YES!  Your business knowledge and perspective will inform and contextualise much of the data science conversation.

Eg: Following the above example, you need to help the data scientist understand the building blocks of customer payment behaviour; ie: what is likely to impact (eg: poor credit rating), where to source it (external credit data) and what is not like to impact (eg: the name of the company).  Company name!?

OK, this is a bit daft, but this is an easy example to find/agree on the engine components, but let’s just the data scientist finds there is a correlation between companies beginning with “S” and bad debt, the data scientist may deem it relevant and include it in the engine/algorithm, but is that right, is that an error. Your role is to add context and help explain what is relevant.

So I can add value now, great! 

But don’t I need to become more data savvy to do data science?

It won’t hurt to learn more about data and the world of data science, but you only need a conceptual level of knowledge of data science and the interpersonal skills to work within the multi-disciplined team.

You need to know the art of the possible, and what everyone brings to the party.  You need to know how to work with data scientist as part of the data project with eyes wide open, expectations set and accepted.
And you do not need to go back to school to learn mathematics, statistics, R, Python, etc…

Eg: You may want to make the final decision on the engine/algorithm/classifier so it can be simplified and explained to the business, “lets only use historical payment performance, as it is a good guide and explainable!” You will want to guide how the data is visualised, as this is what the business sees, and needs to be impactful (after 3-6 months of work!).

And that’s it?

Well, not unless you want to become a data scientist… Besides, you have a tonne of work to do with the business as a value-adding finance business partner, and the above role is highly valuable to the overall success of the project, so do that well, and let the data team do the rest!

Think of this brave new world as learning what your part in data science is, rather than learning data science as a whole.  Zero on that, and you will become the CFO of the future.

Find part two in this series here…