For this edition’s special we have interviewed three ambitious econometrics students that were curious about the applications of econometrics in practice, because: “In theory there is no difference between theory and practice. In practice there is.” – Yogi Berra
Having lots of spare time to fill, Roman Kazus reckoned he would rather spend his time exploring the applications of econometrics than wasting his time watching television. About a year ago, when he got in contact with his current employer during the Symposium ‘Saving Lives’ and had dinner with them afterwards, Roman got more enthusiastic about actually realizing his plans. Being a pianist at a restaurant in Oosterhout – where Roman currently lives with his parents – he was invited to perform at the wedding of one of the owners of the company. The ball started rolling and after the summer holiday, Roman was starting up his laptop at one of the flexible workspaces at the data-driven commercial decision-making company.
The application procedure was quite a happening. He started with a general meeting, which was followed up by a personality test, and another meeting afterwards. The procedure concluded with a skill test, in which programming skills and modeling competences were examined. The test was not necessarily to check what programming languages you already master, but rather to see if you have the right mindset to understand and learn quickly, something which cannot always be abstracted from grades. After the positive outcome of the test, the procedure was topped off with a last practical meeting. This might feel like quite an inspection, but Roman was not the only one that has proved himself. In total, four econometrics students are working at the firm, of which two are bachelor – but soon to be master – students, which is quite a number for a young company.
Methods that are just a little bit more sophisticated than gut feeling or simple statistics already increase accuracy significantly
Roman is putting his expertise in practice for sixteen hours a week and can choose to work whenever and wherever he wants. Nobody is holding you back to keep the work flowing after midnight, at the office, or at home. Flexibility to the limit. The three major positions at the data science company are data engineer, data scientist, and data translator. Roman fulfills the task of a data scientist. In his work, he often employs linear regression and both probit and logit models, which we all have encountered in the course ‘Econometrics’. In addition, clustering algorithms, panel data, forecasting and SQL are used. A lot of statistical programming is done in R, which has an integrated suite of software facilities for data manipulation and calculation, making it more useful to exploit data than, for instance, MATLAB. Since R is also suitable for regression, it is no surprise that the software is newly introduced in the first year of the EOR bachelor program. For the graphical display and visualization of data, they make use of the software QlikView. For techniques that are used frequently so-called ‘blocks’ are coded, which only require input and do the work for you. “This is the most efficient way of working and it is less boring than doing similar tasks ten times over.” Longing for the most efficient process possible, Roman is a good fit in the team. Also, being critical is a characteristic that every member of the company should have. This is because, although projects are handled in teams, all tasks are divided and are mostly tackled on your own. Afterwards, the team meets again and everyone discusses their findings. So, in order to maintain quality, they never deliver something without it being checked or improved by a colleague.
Since most companies know that big data analysis can be extremely useful, but are not aware of the details, the firms’ employees, almost solely econometricians, set up brainstorm sessions with clients to discuss the possibilities of exploiting data. A data request follows, in which all data that might be useful is gathered. Next step: data cleansing. Outliers and corrupted data records are removed, because too much noise disguises relationships. Also, in practice there is no guarantee of significant findings. The biggest eye-opener for Roman was to see that lots of organizations make decisions based on their gut feeling or simple statistics such as averages. Methods that are just a little more sophisticated, for example using double exponential smoothing and including seasonality, already increase accuracy significantly. Now that Roman has caught the data madness, he finds it a good thing that the ORMS master has changed to BAOR and is extended with some more data analysis courses.
For his master thesis, Roman is free to explore the world of econometrics and operations research and can return to the company afterwards. But first, his exchange to Copenhagen will take place.
Tessa was just checking Facebook when a post of her current employer looking for work students caught her eye: perfect timing. At that time, Tessa was doing a premaster and she had two days of spare time she would rather fill with solving problems in practice than in theory. To her own surprise, she, a premaster student, was hired. Apparently there is no need to be a top notch econometrician for a company to benefit from you as a working student. Now, almost two years later, Tessa is still working at the company for one day a week, next to the courses she takes in her ORMS master. To combine the job with her study, Tessa has chosen to spread the ORMS courses over the entire academic year and start with her master’s thesis after the summer holiday. This is no problem, because in practice it seems that combining courses and writing a thesis at a company in one semester is challenging because of overlap anyway.
The firm she works at, a small company located in Den Bosch, is there to create and deliver software for personnel planning and financial and process improvement for those cases where Excel is no longer sufficient. Tessa handles a variety of jobs, from building user interfaces from scratch and customizing and improving existing ones to preparing software training for clients. Her application process went pretty smooth. After submitting her résumé and motivation letter, she was invited to an interview. It was no interrogation, but a proper conversation to get to know each other and learn whether or not she would fit into the team, which seemed to be most important for the company. Tessa did not feel like she was there only to convince them, but also the other way around. No assessment was required during the procedure of becoming a work student.
At the start, she was not thrown into the deep end. Former work student Emmy was still there to show her al the ins and outs of being a work student and explain what would be expected of her. Tessa remarked that most of the activities she does have been learned through practice. She mentions that no specific courses are required, but rather your general intellect and logical reasoning capacity. Nevertheless, some Java and MATLAB programming skills did come in handy. Because of this, there was no need to worry about the company using her own software; if you know the basics of one programming language you are a quick study in the next one.
You will learn what you want and maybe what you definitely do not want
The thing Tessa has learnt most from her job as a working student is how to tackle big projects, from starting a project by analyzing the problem to solving it in a structured way. Furthermore, she learned a lot from working with deadlines that also affect other people instead of just yourself and your grade, which is a whole different responsibility. Since Tessa only works one day a week, time is scarce, so therefore she mostly works on smaller parts of bigger projects. Nevertheless, she really feels like she is useful to the firm, even though she has not graduated yet. She is not just a work student, but really a part of the team. And not only that, she also benefits by getting to know what she really wants to do and would like to do for the next forty years of her life. Piece by piece, she gains experience in consultancy. The only part that she has not yet fully discovered is the contact with clients, which, due to practical reasons, is left to the regular consultants.
Tessa is feeling quite comfortable at the company. Next academic year she will start writing her master’s thesis for four days a week in addition to the one day she is currently working. She felt no reason to look for other companies to write her thesis at, because she already has some interesting topics in mind and she knows she will be given any guidance that she might need. Finally, Tessa would like to encourage econometrics students to go for it. She thinks it is a pity that not many econometrics students apply to become a work student. You will learn what you want and maybe what you definitely do not want, which is okay too. “If you do not feel ready now, you will probably not feel ready after your master’s either. You will have to do it eventually. So why not do it now?”
Valentijn Stienen – Master student
Work student at ORTEC since October 2015
After writing his bachelor’s thesis, Valentijn decided to postpone his master’s; it was time for a road trip through the US and Canada. Next to this, it was the perfect opportunity to gain some experience at a company in the field of econometrics. Curious about how all techniques learned in his bachelor are applied in real-life situations, he just went for it without knowing whether or not – and if so, what – he could provide for a company. The main purpose for him was to get to know which particular area of econometrics would suit him best. “Or not at all”, he adds. “Doing student work and being around full-time consultants is the perfect opportunity to learn and experience what you can provide a company and to discover your future interests!”
His company sells optimization software and analytical solutions to increase the effectiveness of business. What business field in particular is not important; from the manufacturing sector and retail industry to the sports sector: “There is always something that can be optimized. The wide variety of problems where econometric techniques can support is enormous!” The application procedure was a little more sophisticated. As is often the case in applying for consultancy work, an assessment was required to test Valentijn’s intellectual capabilities. Once he passed, everything was as thick as thieves: he was invited for a relaxed and informal conversation at the office so that the firm could find out if he would fit in their team. “In the end, this is more important than having the highest grades of your class.”
As he is now working in the analytics and optimization team, Valentijn mainly focuses on the analysis and visualization of data. The latter is of great importance, Valentijn stresses, because they want the company to understand and be involved as much as possible. So, in addition to the data analytics, he also creates dashboards that show new insights in data. At the start, Valentijn was given time to develop skills in the used software. After he mastered those skills, he was assigned to a project in which he could actually use them. Indeed, team members actually benefited from his skills, as regular team members were not ashamed to ask if he could help them out. Because of this, Valentijn really feels like an additional member to the team and not “just” an intern. Throughout the process, guidance was provided by the project manager wherever needed. Fortunately, quite some techniques from class are applied in practice, such as regression, linear and combinatorial programming and game theory. Also, the mathematical way of thought in programming, but not the programming language as such, has been useful to Valentijn.
The wide variety of problems where econometric techniques can help is enormous!
This brings us to Valentijn’s main learning point during his student work: the experience of working at a real-life OR company. In practice, problems are not clearly stated on a piece of paper and are supplied with corresponding data. “You should operate differently to get a job done than in assignments of courses. Defining the problem is already done for you in your class. In practice, it is not.” Instead of having it provided to you on a silver platter, you should gather your own data. Another concept that is crucial for the quality of your end result is deciding what technique to use. “In class, the assignments are subject specific and are based on the content of that particular course. Hence, the appropriate approach is clear right away. In practice there are many ways to go. It is up to you to choose which way is the best.” Only after that you start modeling, solving, and implementing. Additionally, teamwork and communication come into play. “In practice you always work in teams, which is great, because you derive new insights by brainstorming with your colleagues.” Valentijn states he has learned how to explain everything efficiently and clearly, both to his colleagues and the clients, but with enough detail for it to be useful.
As of September, Valentijn will start with the newly introduced BAOR master. If the travel time from Tilburg to Zoetermeer is manageable, he would like to stick with his employer during his master’s thesis. The sixteen hours a week he has worked since October 2015 have helped him figure out that data analytics is absolutely up his alley. Nevertheless, Valentijn would rather switch between fields once in a while than sticking with one particular expertise of econometrics. Other fields he would like to explore in the future are for example finance and operations research. From experience, Valentijn would recommend everybody to follow an internship or work as a student assistant at a company. “It is a great experience and no time is wasted! If you have some spare time, go for it!” And if you are too busy, he would suggest you to create spare time, even if this results in study delay. “Only the experience is already worth it.” Secondly, feel confident! “Do not forget that this is also an opportunity for the company, since they might have just harvested new talent! Be yourself. If they do not hire you, you would probably not be happy to work for that particular company in the first place.”
Text by: Steffi van den Hanenberg