– Most university projects are submitted digitally and after some time they are simply deleted from the disk. Therefore, the greatest advantage of preparing posters was the ability to see some real results of the entire semester's work – says Jakub Horodyski, a student of Information Technology in Automation Systems.

Dr. Agata Migalska from the Department of Control Systems and Mechatronics runs the course "Methods of data analysis and extraction." In order to complete the course, students of the second year of master's studies in the field of Information Technology in Automation Systems are required to prepare a project. In groups, usually consisting of three people, they worked on analysing phenomena and problems that they observed in everyday life. For this purpose, they had to use data analysis and knowledge extraction techniques. Last semester, Dr. Agata Migalska offered students a different form of assessment than usual. They were supposed to present the results of their work on posters. It was a perfect idea.

– In my opinion, completing the project by preparing posters was a great idea. I deal with computer graphics as a hobby and I was very pleased with the fact that I had the opportunity to prove myself – says student Jakub Horodyski. – A competition for the best poster organized by Dr. Migalska additionally motivated students to work. – Most university projects are submitted digitally and after some time they are simply deleted from the disk. Therefore, the biggest advantage was the ability to see some real results of the entire semester’s work.

Rafał Szyszczyński, a student of the same field, has a similar opinion: “The form of assessment is an interesting idea and a nice break from classic reports. In the case students encounter any problems in their research, Dr. Migalska is also willing and ready to offer assistance and suggest other strategies to obtain better results. I can confidently say that new students will also like this form of classes.”

IT specialists more susceptible to depression

The scenario was simple: students chose a problem they would like to solve. Then they had to find the relevant data, analyse them and solve the problem. Sometimes the problems were not clearly defined, the group implementing the project simply wanted to see what conclusions they could draw based on the data.

One group focused on students' mental health. She researched the causes of such conditions as depression, stress and anxiety. It turns out that, among others, the field of study is very important and that computer science students are particularly susceptible to depression. The data concerned American students, but it can be assumed that they do not differ much from the situation of Polish students.

Another team analysed data on city bikes and when people rent them. Effects? It is hardest to find a free bike on Wednesday, because that is when most people rent them. The least people rent bikes on Sundays. The students confirmed what intuition tells us, that the weather is also crucial – the warmer it is, the more willing we are to use the bike.

There is also more scientific work: students investigated which type of neural network is better for weather prediction. It turned out that the best network is such that predicts all variables at once, instead of n models that do it for each variable separately. – The group implementing the project chose a very specific model and checked two variants – adds the leader.

Jakub Horodyski and Rafał Szuszczyński dealt with diamond prices. 

– Our team wanted the project to be as practical and real as possible. The topic of our work was the prediction of diamond prices according to their most important features, so the goal was to develop an algorithm allowing the valuation of a stone following its appropriate parameters – says Jakub Horodyski.

– Our group was looking for a large database that would meet the guidelines presented during the classes – explains Rafał Szuszczyński. – We particularly liked the diamond database because we thought it was an interesting issue. Moreover, it fits into the area of our direction based on programming models that try to predict the result based on the provided data.

Both students admit that the results were what they expected. Rafał Szuszczyński explains that the inaccuracy of the prediction is the result from different prices for the same parameters of diamonds. „This was the result expected by our group" – says the student.

Women focus on their partners' beauty

What topics did students research? The topics were very diverse. One of the projects concerned musical tastes. Students investigated whether there is any connection between what we listen to and, for example, Blue Monday. It turned out that there is no such connection – we do not listen to depressing music on this day. However, on holidays we listen to music that is different from our preferences.

One group decided to look at data about people who speed date. Speed dating participants answered three groups of questions. In the first one, the respondents talked about themselves, what they were interested in, what they were looking for in a partner, and what features they considered important in him/her. Then the dating organizers asked what they expected from their partners, whether they had to be attractive, whether they had to be cheerful, or whether they had to like animals. The third group of questions concerned the results of speed dating – whether the respondents believed that they would have a happy relationship with the person they connected with. Conclusions?

– Students examined how we think about each other within gender. Surprisingly, women are more likely to expect their partner to be attractive, while men say they have no such expectations. The students also developed a model that predicted whether a couple would continue dating after the speed dating experiment ended – says Dr. Agata Migalska.

Sandwich method, i.e. group feedback

In addition to substantive support, the teacher also organized a session for students where they could receive feedback. It was also an opportunity to practice soft skills.

– I supervised the process and gave instructions on what to do next with this or that model to obtain a reliable result. Later, the students made posters, but before they printed them, we had a session during which each group presented their poster, says Dr. Agata Migalska. –After the presentation, students gave each other feedback on what else could be changed or improved to improve the poster. After this session, they had a week to apply corrections and they actually refined a lot of details.

The effects of the work of Dr. Agata Migalska's students can be seen at the exhibition in Technopolis on the first floor.