For this edition of ‘Into the Business,’ Estimator travelled to Amersfoort to interview Paul Rooijmans, founder of data company Lynxx and Simon Langbroek, a graduated econometrician that is currently working for the Lynxx as a data scientist consultant.
Could you describe the structure of Lynxx and the type of people who work there?
Paul: At the moment we have approximately thirty colleagues of which five or six are working for us in Sydney. We would also like to open more offices in London and Berlin. People working here usually have a quantitative background. Most studied econometrics, but some also did their bachelors in mechanical engineering, mathematics and physics. While we value a mathematical background very much, we also acknowledge the importance of social skills. When someone submits an application, we analyse how this person fits into our team. For that, we allow himor herto participate in a case.
Simon: Lynxx has three working fronts, all mainly within the public transport industry. Firstly, we help our clients during big tender projects where our clients are bidding for the right to facilitate the public transport for a certain area for a number of years.. We help these clients by performing data analysis/modelling in order to optimise certain processes, calculate effects of certain initiatives or get insight in specific aspects. Currently, this service is available in the Netherlands, England and Australia. Secondly, we perform business intelligence analysis. We provide insight into different processes based on data and Key performance indicators (KPI), creating a tool with which they can improve their organisation. Finally, we build statistical and machine learning models for our clients, using data to make for example future predictions. Generally, we analyse what is correlated, and what this affects to get the best understanding of the situation.
Inside the office of Lynxx
Paul, how did you become one of the founders of Lynxx?
Paul: I have a background in public transportation, and eight years ago I started working for a small consulting firm. This firm was responsible for introducing the OV chip card, and consequently for closing all ticket windows. While working on this project, I realised how much data was circling within public transportation.
Simon, how does a regular day as an econometrician look like for you?
Simon: It differs. Sometimes I am in the office all day. Other times I spend my day with clients, defining their problems or presenting our work to them.
Do you work on projects on your own?
Simon: Most of the time I work with colleagues. It is less fun to work alone, and we can also learn from each other.
Paul: It is important for colleagues to work together because each of them has different qualities. While one may focus on content, another can focus on design. Both are equally important.
Can you describe the process of getting the big data from customers up until the final acquisition of applications and useful information?
Paul asks us what a problem at EUR is; we answered "the busy university library (UL)."
Simon: In the beginning, we have a conversation with our client to try and decide if the problem is a problem. If that is the case, we also try to define it. For instance, in the UL case, we have to ask ourselves, "do students feel like the busy University Library is a problem?" Then we decide if data analysis can help with the solution. Further, it is necessary that we know what type of data is available. For UL, that would be the number of Wi-Fi logins, coffee machine usage, computer logins and how many times the electric doors open. Finally, we get insight into the data. That would be something along the lines of who visits the UL when it is busy. After this, we decide if the results match our expectations.
Paul: Finally, we ask ourselves if it is worth it to solve the problem, and if that is the case, decide to proceed. For UL, that could be a dashboard on which students can see if it is busy. We can always test multiple solutions.
Do you spend a lot of time cleaning data?
Simon: It depends on how the client delivers the data. Sometimes we are lucky, other times it takes a long time before we can use it. Other times, there is no match between the data and the problem.
On your website, you describe yourselves as "the people who look at things differently." Could you explain what that means?
Simon: We look much further than other people by extracting information from data most people did not know exist.
Paul: For example, an electric bus generates so much data from which you can extract information.
Simon asks us what data is generated by that bus; we answer the number of check-ins and how often it needs to wait for traffic lights.
Paul: That is just level one. Level two would be how often they replace screen washer fluid. This provides information about how clean the environment is. Further, we could also compare the number of check-ins and weight on air suspension, understanding more about the overall population weight in a neighbourhood. I believe the combination of creativity and intelligence is the key to success.
What project are you proudest of?
Simon: I have recently worked on a dynamic simulation model. In python, I made a model in which you can simulate the operation of a public transport network. You can analyse the effects of decisions on for example punctuality. This model has been very successful. The client could evaluate the effects of decisions like for example investing in an extra platform to improve punctuality. This is the type of problem that benefits from data analysis.
Paul: I am just proud of the current team, and how we can help clients save money by optimising their services.
Which added value do econometricians have in comparison to other quantitative professionals?
Simon: This value is less relevant than I imagined before I graduated. An econometrician has more experience with machine learning and a strong mathematical and statistical background. However, most people with a strong quantitative background can learn the rest very easily.
Paul: We find econometricians to be a very good match with our company. It is easier to teach an econometrician how to program than an IT professional about statistics.
What are Lynxx's future ambitions?
Paul: We intend to grow the company significantly in the incoming years. I think we are going to be five times as big in the next three to five years. We would also like to bring in new people two fulfil out three main areas of action. We need to make sure we get the most qualified people involved. As we previously mentioned, we would also like to open new offices in different countries. Finally, we would like to further our company's creative aspect, by challenging us with new projects.
Do you have some advice for our readers?
Paul: Do something that stimulates your creativity, such as travelling or helping out at a theatre festival. Bringing yourself outside of your comfort zone provides you with new perspectives.