“We have to break out of the comfort of accepting things without questioning them” – Data Scientist Nadja Keidel
In this interview, Nadja Keidel, Expert Data Scientist at Zühlke Engineering AG, talked on how she became a coder, that there is a major difference in ‘coding’ and ‘programming’ and why we don´t have to be afraid of the future.
Nadja, you are a data scientist and an R & Python coder for the Zühlke Group. How did you get into coding and what makes it exciting for you?
The first time I came into contact with code on my data scientist career path was during an introductory lecture as part of my mathematics degree. We had to write a program that would convert the classic ‘Hello, World!’ message into ‘Hello, Nadja!’ Everyone starts small (laughs). During my degree, the focus was on R & MATLAB. Python was added into the mix later on.
Among other data scientist skills, I went on to learn that there is a major difference between ‘coding’ and ‘actual programming’. Or, to put it another way: code is a program that is running. Actual programming requires much more input, as it consists of a readable structure and well-implemented logic. You tailor your approach based on best practices such as clean code and work together to show, review and learn how to improve the codebase – something which is important in IT in general. This is something I learned while working as a data scientist in the industry, particularly when swapping information with the software developers at Zühlke.
I find it fascinating that being able to program opens up an entirely new world to you – it’s like learning a new foreign language. You suddenly understand things that were previously concealed from view. What’s more, programming is a discipline where you see the results immediately and that’s important for the work of a data scientist.
There’s also a certain beauty about a line of code. It’s like in mathematics: a good proof is clear and understandable, and you have your own little ‘Eureka’ moment. You experience this sensation in programming, too, which is both fun and requires you to be more creative than you might think. Again, it’s like learning a foreign language: nobody is keen to read a text consisting solely of long, complicated sentences. We data scientists like to see clear, logical statements.
One of the fields you focus on as a data scientist is machine learning, quite an exciting subject, which is followed by many questions about the future, e.g. machines replacing some occupations. In your opinion, how will AI change the professional world?
As I see it, we’re right in the middle of a sea change right now. And like every new technology, the initial focus is on exploring all of its various limits. In the process, we learn what works and what doesn’t. I think the greatest challenge lies in changing the way we approach the subject. In using machine learning algorithms, we are leaving behind the world of absolute numbers and entering one of probabilities – a crucial turning point on my data scientist career path. Up to now, we have been able to rely on a computer giving a right or wrong answer. This is no longer the case, as a machine learning algorithm does not depict an absolute value and is never 100% correct.
This is why we need to be more critical in our thinking, and those working with these technologies need to have a great deal of accountability. I believe it is extremely important that we face up to this responsibility and actively create the transparency that is required. Everybody should have a basic awareness of how machine learning works – and what it is not capable of doing.
Nowadays the future is often described in a negative manner, there is a lot of uncertainty about future technology. As a data scientist, can you offer some reassurance?
Behind every technology is a host of different options for using it, and behind those options are the people who can use the technology in a meaningful way. The worst scenario I can imagine as a data scientist would be if the knowledge about a certain type of technology is entrusted to a small group of insiders. Knowledge is power – that isn’t going to change any time soon.
I want to build on the previous statement here: we have to break free of the mindset where we look at things without questioning them. This is not a radical new concept. Back in history class at school, for example, we were told to always check and support statements and facts based on multiple sources. Data science requires you to do the same. Critical thinking, a healthy amount of scrutiny and not blindly accepting everything as the truth: these are already the key pillars of an enlightened society – and this will continue to be the case in the future and when implementing new technologies. Such an approach will enable us to benefit from the many advantages and opportunities provided by these new technologies.
How can AI improve the future of humanity and what are some sectors where this is already the case?
Assuming that they are used sensibly, these types of technology could present a huge opportunity in, say, medicine. We have the chance to benefit from highly personalised medicine, which from the patient perspective is certainly desirable. If I become ill, I can receive treatment that is tailored exclusively to me and my body. Other examples include projects that use algorithms to scan X-rays to identify certain disease patterns or to make triage simpler for overworked doctors in the event of an emergency.
There’s also a current topic in journalism which is really exciting from the perspective of a data scientist: Associated Press (AP) is now creating automatically generated reports, such as for business performance (quarterly figures) and sports results. This enables AP to use AI to publish reports on various companies and marginal sports, which makes it easy to inform readers about how their favourite team did in the fourth division. Previously, you would have had to assign a reporter to something like this for half a day. With less routine work to do, journalists can use their time to tackle important topics through investigative journalism and so on. The idea here is that people can be released from their dull routine work and focus instead on exciting or important issues. This desire isn’t confined to any one industry.
In addition to being able to address exciting, important topics, this could also be our chance to benefit from more free time. What could we achieve with the time we save if AI takes care of many or all of the routine tasks we’re currently saddled with? Perhaps society will change to the extent that we no longer have to spend as much of our time working as we do now. Apart from data scientists.