Being the only Data Scientist in your company: Interview of Thomas Dolman

Thomas Dolman is a Data Scientist at RegioPlan. What makes his profile atypic is that he is the only Data Science professional in his company.

Therefore, his work is much more entrepreneurial in the team. In addition to doing proper code, Thomas brings data enthusiasm in his company in order to implement a data-driven culture.



Can you give us a brief introduction to your professional background?

I attended the University of Amsterdam where my educational background was Neuro Economics, which is a field that seeks to explain human decision making and the ability to process alternatives and decide on an action. It was a very challenging master's degree dealing with large quantities of data, at different levels of complexity. The course had a heavy focus on Data Science, which is where I realized I was interested in working with Data.

I then joined a consultancy firm as a data science consultant. There I worked in a relatively large team. I worked there for 9 months before realizing that consultancy was not really for me due to the sole focus on financial gain and not societal improvement.

I decided to make the real switch and ended up in RegioPlan, where I have been working for a year now. With RegioPlan, we work mainly for governmental organizations, municipalities, and administrative. We research policy and give policy advice. As a Data Scientist, I implement data strategies. It means I implement decisions based on Data but I also introduce new concepts.

Can we have an impact not using questionnaires or by interview but instead through web scraping and social media analysis? On the reporting side, we do visualization and dashboards to make policy research and reports. It’s exciting to do this at RegioPlan because they have been experimenting for a long time now. The company has been running for about 35 years. Back in the day, they attemptet questionaires on floppy disks  which was a really large innovation at the time. This goes to show that the company has a rich history in implementing new strategies and assessing what fits and what doesn’t fit.



What are the big projects you have been working on since you arrived at RegioPlan?

It is very diverse because I’ve made data parts on several projects. Some are for governmental organizations and some are not. I’ve been focusing on 2 educational projects where we monitored government agreements between schools and ministry.

We have made data analyses on that. I also did text analysis on zoning plans. In which municipalities make agreements on what a region should look like, type of buildings, height, greenlands, etc. We do text analysis to assess which type of studies have been done before the development of the zoning plans. This is used to estimate the cost of a new zoning plan. 

My responsibility is to find new ways to do these analyses, to show this and to propose it to the rest of the team. Then we decide if it is valuable or not for company activities. So it is very dynamic and there are a lot of projects to work on. We also have to bring a lot of expertise to it because we want to check the accuracy of the numbers we are spreading out. What is also really important is being able to explain the actual models to the team and the clients. As I am working for governmental organizations, we give simple models that are easy to explain such that citizens can follow the decision process. 



You’ve worked in a large team of Data Scientists before. How have things changed in your daily work now that you are the only Data Scientist at RegioPlan?

There are a few differences.

There are a lot more opportunities to try new things because you are constantly working with people with great expertise in the public policy domain that don’t know about data science. We are asked to work with data-oriented processes so my work has to get the interest of the decision makers like: “Oh, I like the thing that you implemented. I could do this and that with it.” It is similar to my previous job in the sense that we work in sprints. We deliver the work in short term periods and we show people what we could do with that.

What I do here is pretty different from what I did in a Data Science team. Before, my goal was having good code, quality checks and versioning (this is really data science team-oriented work). But we don’t sell our idea at the end. It was much more like my code was the end-product.

Now my end product is that my colleagues are coming to me and say: “Hey, 2 months ago you did that, and I want to do that again.” First, you have to be enthusiastic about that, think about what is possible and then you give them certain tools and they start to think differently about possibilities and different solutions. In that sense, it is far more entrepreneurial and you have to be more structured. When you come back to your colleagues, you have to get feedback from them in order to improve your solutions.

“What’s wrong? What could be improved?”

Also, with my team, I don’t want to enter in technical discussions on “text mining” or “natural language processing” because these are subjects more inclined to data professionals. Here, I come with the direct benefits of my work and I demonstrate it with words that are understandable for everyone. Then, my colleagues can come back to me with requests to know if they could use my solutions for their work.



How does your team view your role? How do your colleagues relate to you?

I think that my team has a positive idea of Data Science since they are very focussed on innovation.

There is a lot of interest during conversations and meetings and if someone sees it as a technical subject, I always try to make it more comprehensible for them. It is also a practical thing that you can use and obviously people see this at different speeds and it is actually very interesting to decide how we can utilize data to its full potential.

I also managed to create and feed enthusiasm around data science. When you start as a new Data Scientist in an organization, nobody knows what you can and cannot do. During that initial time, you have to feed them with ideas and propositions.



Concerning your data projects, how do you manage your projects? How do you communicate on the advancement of the tasks?

Usually, my team asks me a question regarding a research topic, then, I pretty much have to figure it out.
Being the only Data Scientist in a company also means that there is no one that you can ask for coding advice. You have to figure it out and if it is difficult, I reschedule the deadline for the data science work that I do. You have to be able to look for a simple answer when you have a problem. Initially, there is a lot of help on the online data science community side. Thankfully, there is a large community available. 

When you start to become more knowledgeable in Data Science, you start to understand that you don’t have to solve every problem. I can also connect with my expert networks and ask them if they have ever experienced my difficulties.

Especially when I want to know which direction I should choose. I usually start by asking people around me what they would do.

But I think it is down to personal character, you have to be ready and creative when you need to find solutions. So yes, it is easier and more comfortable to do it in a team but when you do it alone, in the end, you know how it works because you figured out how to manage it by yourself so the learning curve is much more exponential. You also have to be conscious and accept that you will not find the most optimal solutions because I do think that you need a team for that.



What experience and skills do you need to succeed as the only Data Scientist in a team?

For me, the first thing that will make you successful as the only data science person in a team is the personal attitude. You don’t want to be skilled in only coding and programming languages. You need to be skilled in communication. You need to have an entrepreneurial communication where you come with the direct outcomes of your projects.

You also need to be sure about your data and it is hard to do programming checks with your own data. But you need to do that because, in the end, it has to work properly. You need to be broad and ready to find practical solutions quickly. 

Finally, you have to be ready to confront people that are not interested in Data Science. You'll have to create and lead interest. It can be demotivating so you have to be ready for that.

You can change your personal attitude through different events and causes. For me, my previous experience in a Data team helped me to be far more entrepreneurial, pitching and selling more.

Today Data Science is changing the way of making decisions so you also have to implement new rules and changes on the managerial aspect. You need to lead your decision makers through a more data-driven mind. You will have to change their way of working and I think you need a bit of arrogance to do that.



What advice would you give to a Junior Data Scientist that starts working in a company as the only data science person?

Be very aware of the organization you are entering in. You have to like the impact that you will have. You can do business implementation so you need to be sure that you will love that impact. You have to like the contribution that you will have because it is what will motivate you. Then, I would say to meet everybody and not only shake their hand once or twice. It’s having a meeting with everybody - decision makers - to clarify your plans and aspirations. The last thing is getting a lot of feedback from your colleagues and team.

When you get the big picture of that, then you’re ready to implement best practices.



About Regioplan: RegioPlan helps governments, non-profit institutions, sector organizations and companies with research, advice, and secondment.

Follow Digital Source on Linkedin and read our daily articles.