Waseda WeeklyWaseda Weekly

"I want to study data science more and improve myself."

Machida Hayato, 3rd year student, School of Political Science and Economics
Matsuda Takahiro, 3rd year student School of Political Science and Economics

The four members of "Team Tanuki" at Building No. 121 on Waseda Campus. From left: Saeki, Matsuda, Machida, and Nagase

​The Waseda University Data Science Competition (hereinafter referred to as the Competition) has been held annually since 2019 as a learning opportunity to learn data science in a practical way. The sixth competition was held in November 2024.Under the common theme of a data science orientated approach to interests and problems discovered during time at Waseda University, " 22 teams each set their own objectives and challenges, and competed on the effectiveness and novelty of their analytical methods and the knowledge they gained. The winners were Team Tanuki, consisting of four members of the seminar of Associate Professor Toyama Yuta (Faculty of Political Science and Economics): Saeki Reika, Nagase Satoshi, Machida Hayato, and Matsuda Takahiro.  We spoke to Machida and Matsuda, who represent the team, about the process of winning the Grand Prize and their future prospects.

--Please tell us how you came to enter the competition.

Machida: In 2023, two teams of seniors from my Toyama seminar participated in this competition, and both teams won prizes. Because of that, Professor Toyama suggested that our year enter the competition this year, and we decided to participate. This year, eight third-year students from the seminar split into two teams and participated.

Presentations on the day of the competition. As a result of the pre-screening by video and the poster presentations on the day, the seven teams selected were called upon at the venue and gave presentations in front of all the other teams and the audience

--What made you decide to join Toyama's seminar?

Matsuda: When I entered School of Political Science and Economics, I learned theories such as microeconomics and macroeconomics from my first year, but personally, when I heard about these theories, I would wonder, "What does it actually mean?" and I never found them very interesting. There are many seminars that analyze empirical data, but I was interested in Professor Toyama's specialty, industrial organization theory, also known as the economics of the enterprise, and I wanted to join Toyama's seminar.

Machida: I also took Professor Toyama's class on industrial organization and found it interesting. In the Toyama seminar, where we study empirical microeconomics, we conduct empirical studies to back up theories with data. I was fascinated and interested in the idea of being able to use data to prove theories that make you wonder if they really have any meaning in society.

Photo on the left: Group photo from the last seminar of the spring semester of 2024. All third and fourth year students make a T-pose in honour of Professor Toyama
Photo on the right: September 2024,  A scene from a seminar camp held atKaruizawa Seminar House​

Matsuda: On the other hand, I learned about data analysis by taking classes offered by School of Political Science and Economics and the Data Science Center. I also learned how to do simulations by myself while actually conducting analyses.

-What was the content of your winning presentation?

Machida: The theme of this competition was "Analyzing interests during time at Waseda University," so we first picked out topics related to everyday student life that could be used as themes for analysis. Various themes were raised, such as the Big Six University Baseball, abura-soba noodles, and revitalizing shopping districts, but we decided on the analysis theme of "Proposing new rules to increase attendance at Tokyo Big Six University Baseball," as this was something that was easily accessible as open data, and our team member Nagase is a big baseball fan.

Photo on the left: Nagase and a friend went to watch the Waseda-Keio game
Photo on the right: The trend in attendance at Tokyo Big Six Baseball games from spring 2018 to spring 2024. From the competition presentation materials (click to enlarge)

Matsuda

Matsuda: The objective was to "improve the current points rule to see if it would increase the number of spectators." The current points rule is a round-robin tournament between six teams: Waseda, Keio, Meiji, Hosei, Tokyo, and Rikkyo. The first team to win two games against each university gets one point. Therefore, in addition to the maximum number of points being five, if a team wins one game but loses two, they cannot get any points. Therefore, we thought that if the points are lost, the expected value of winning drops drastically, and the dynamics of the ranking changes will disappear, which is an issue. Then, using a method called multiple regression analysis (※), we clarified how much factors such as ranking volatility and the possibility of winning affect the number of spectators.

(*) This refers to a type of regression analysis that has multiple explanatory variables. Regression analysis is a model that shows how a different variable changes in response to the change in another variable.

Machida: We proposed a new rule that would award one point for each win, and when we ran a simulation under this rule, we found that the average attendance would increase by 200 people. An increase in attendance not only leads to increased revenue and the promotion of university baseball, but is also expected to increase student interaction, energize the six universities, and boost player morale.

The results showed that the rule change would extend the chances of winning and increase the opportunities to gain points

-Did you have any difficulties when taking part in the competition?

Machida

Machida: The purpose we had in mind at first was not to propose new rules, but to roughly understand the mechanism by which the number of spectators is increasing. So we analyzed the data from various perspectives, such as the relationship between the travel time from each university to Jingu Stadium and the number of spectators, but even after a month or two of trial and error, we couldn't get any interesting results... I wondered what to do, and when I discussed it in my seminar, a senior who had been helping me from the early stages of the research suggested, "Why don't you try changing the theme?" With only a month left until the presentation, we all struggled over whether to restart from scratch or continue as we were. Even so, Professor Toyama told me, "One option is to persevere and keep trying," so we decided to see it through to the end.

After that, another senior colleague came up with the idea of "changing the rules," so we recollected the data and analyzed it under various simulations. In the end, we were relieved to have produced some interesting results, and we are really glad that we won the top prize.

-Finally, please tell us about your future prospects.

Machida: I would like to utilize the knowledge of data science and economics that I have cultivated so far in a professional field. Currently, I am doing data science in R, but while I was looking for a job, I learned that many companies are doing analysis using machine learning in languages such as Python, and I felt that I still had a long way to go before I could enter the workforce. In the future, I would like to study more data science, including machine learning, and improve my own level.

Matsuda: On the other hand, I'm not really thinking about becoming a data scientist or anything like that. Rather, I think that data is probably used in every company, so I hope that I can utilize the data science I've studied so far in the job I get.

Saeki and Nagase from the team were also present during the interview

No.891

Interview, text and photography: Waseda Weekly Reporter (SJC student staff)
Nishimura Rinka, 2nd year student School of Human Sciences