Leonie Finke (grade 11) reflects on the recent collaboration between Roy Tomlinson (AP statistics teacher), the Japan Center, and GE Healthcare Japan to take students to a manufacturing plant and experience real-world statistics issues.
Last year, Roy Tomlinson (high school teacher) collaborated with Mariko Yokosuka (Japan Center co-Director) to enable a small group of AP Statistics students to visit the GE Healthcare Japan’s manufacturing plant in Hino. The collaboration between students and GE went extremely well and Mr. Tomlinson set out to offer the same experience to this year’s students. And “experience” is the best word to put it because this event was absolutely mind blowing.
One week before the visit was scheduled, students received a massive amount of data from GE. The problem: the business observed increasing volume of defective ultrasound probes at the final test. Students were told to analyze the data and identify the factor that was most strongly correlated to the outcome in the final test. The amount of data we received was, for many, overwhelming. In the past, we had worked with up to five categories and with far fewer data points. Now, we were presented 14 categories and over a thousand data points. Mr. Tomlinson commented that, “I think it was great for our students to look at this large data set and try to apply it to real-world situations. Our students did a great job of putting all the ideas we had learned in the class together.”
Kaitlyn Wu (grade 11), who went on the trip, explained how “When you are just starting out and learning new concepts in statistics, the data given to you is usually structured so you can find a reasonable answer through standard steps. However, when working with real data there are many factors you have to consider and lots of trial and error.” Hikaru Hayashi (grade 12) described analyzing such a large amount of data in order to solve a problem as a “full circle moment in our AP Stats class.”
It was difficult to understand the meaning behind the data. We had been given a basic diagram of how the variables were connected chronologically, but had no details on what mechanism each variable referenced, and were given standardized data. Though this did make it more difficult to make logical assumptions on the relationships between the data, it also mirrored more of a “real-world” experience as third-party organizations often have to deal with standardized rather than raw data in order to keep the inner workings of the company private. Still, we faced the challenge head on and worked vigorously to solve the issue at hand. Analysis methods including comparing distributions of data for different categories using histograms and boxplots, using chi-squared goodness-of-fit tests to evaluate whether the differences in the proportion of failures within one variable were significant, creating linear regression lines between variables in order to determine whether there were any strong correlations present and controlling different variables by only examining data within one category of a variable.
On Friday, May 25, students, accompanied by Mr. Tomlinson and Kristi Hoskins (high school teacher), made their way via train and bus to the “Brilliant Factory” in Hino, Tokyo. There we were introduced to Juro Kawakami, a data scientist and project manager who has worked almost 30 years at the GE Healthcare Japan’s plant in Hino in the statistics and engineering departments. In a lecture by Mr. Kawakami, we learned about the formation of GE and its branch in Japan.
Mr. Okuyama, another data scientist and so-called “Mr. Scrap Reducer,” joined Mr. Kawakami in giving us a tour of the factory. The basic diagram of the manufacturing process that we had received began to make a lot more sense as we learned how each individual mechanism was carried through, overseen and subjected to the collection of data. As a part of the “Brilliant Factory” initiative, the Hino Plant has become a model for GE plants around the world.
After finishing our tour, it was time for us to present our findings. Preparing for the presentation had been, at first, slightly chaotic. Because we were all in different classes, there was initially very little communication between groups. However, as the day of the visit drew closer, we learned how to effectively coordinate. Actually presenting were Aishwarya Kumar (grade 12) , Bar Savion (grade 11), Tomoki Yamanaka (grade 12), Kaitlyn Wu, Sara Schwartz (grade 11), Hikaru Hayashi, and myself. We tried making our presentation a mélange that was comprehensive for both professional statisticians and representatives with only a basic understanding of statistics. A big score for us was when, in the middle of our presentation, Mr. Kawakami paused and stated that he had learned something completely novel from our analysis. One of our team leaders, Bar Savion, said, “I think we did really well in our presentation, and even with the short amount of time we had, we were able to articulate a well thought out presentation with correct conclusions. One thing we could have done better was …[to] see whether the conclusions we made are actually applicable.
In response, Mr. Kawakami explained the actual problem, and what we might have missed. One remarkable comment was that, when including the importance of chronology in analyzing the data, new patterns could appear. Another difference between the students’ and the GE team’s analyses was the program used. While we used TI-nspire and Excel for the majority of our graphing and calculations, the GE team used a data analysis JMP, which, when it came to visualizing data, was a game changer. Overall, however, our findings were amazingly insightful considering the restrictions. We then proceeded to do a group exercise where we received more data and had to work together to analyse the data and come up with a plan for improvement.
Finally, we shared a bento-box lunch with representatives of the data analysis team. The GE Staff were energetic, warm and attentive. Speaking with professionals about their experience was incredible, as it gave us a new perspective and showed us how much statistics was really applied in the professional world. Junior Samantha Walker described how Koki Yasumoto, a data scientist who specializes in artificial intelligence, showed students a program that he had recently written which, when implemented, would increase factory productivity.
Reflecting on the event, senior Kai Weiss said, “Taking AP stats, I had always assumed I had a reasonable understanding of stats that could applied to the real world, but working with real data made me realize there were a lot of things I still didn’t know.” Junior Rei Lindemann noted that “it’s great that we get to see stats in action and engage with real non-textbook problems. More classes and subjects should do things like this.”
I was blown away by this experience. I took AP Statistics because of my interest in pursuing a career in science. When I enrolled in the class, I really didn’t have any idea what statistics actually was. Over the past year, I have learned a staggering amount of information about data and how it is can be examined and manipulated. This trip, in particular, was truly invaluable because it showed me the true importance of statistics in the modern world (and, honestly, also resulted in me having a huge nerd-out). In 2018, 2.5 exabytes of data are produced daily. After taking this class and experiencing the application of it in the real world, I would recommend everyone to at least take an introductory course in statistics.
We would like to thank Ms. Yokosuka, Mr. Tomlinson, and the representatives from GE Healthcare Japan for providing this amazing opportunity.