I am very proud to announce that i won the “Teaching Award 2014” of my university!
My teaching concept “More Interaction with Students thanks to Flipped Classroom” was selected to be the best and most convincing one. Wow.
The award comes with a prize money of Fr. 5000,- which I will use now for a project on innovative teaching.
Before i start, let me just explain what a “Flipped Classroom” is (you may skip this paragraph if you already know the concept). Flipped Classroom is a teaching method where lecture and homework are “flipped”: First, students read part of a text book at home; then they solve some simple online questions about the topic; finally, in the lecture, they work with the teacher to clarify open questions, discuss the topic and solve exercises. For more details, see e.g. flippedlearning.org.
I have been using “Flipped Classrooms” for a while in my lectures, and feedback of my students on this teaching method was always very positive. They pro-actively participated in discussions, presented their solutions, over 90% prepared their homework, and they scored well in exams. However, i was never sure if the method really “works”, if it is – at the end – any better than classical teaching methods.
For this reason, i initiated a research study on the effects of Flipped Classrooms. The study was conducted in 2014 by Geri Thomann and Andrea Keck Frei from Zurich University of Teacher Education, on behalf of School of Engineering of ZHAW [1]. In the study, three classes of computer science students were compared, one of which was taught with Flipped Classroom, the other two with classical lectures.
Although the sample of the study was rather small, the results are very encuraging. Here are the main results:
Overall, the study concludes with “a positive conclusion for Flipped Classroom” ([1], translated). If you are interested in more details, I have prepared a concise summary of the study in [2].
[1] Andrea Keck Frei und Geri Thomann: “Begleitstudie Flipped Classroom ZHAW Informatik (Ergebnisbericht)”, Zurich University of Teacher Education, October 2014.
[2] Mark Cieliebak: Auswirkungen von Flipped Classroom auf Fachwissen und Kompetenzen von Studierenden, Zurich University of Applied Sciences, December 2014.
“Can you find and visualize some interesting insights about garnment industries within one hour?”
That was the challenge we were give at my first “data expedition”, which i joined while i was at okcon (an international conference on open data).
The concept of a data expedition is simple but effective: bring together a small group of smart people: designer, analysts, story teller, researcher, IT guys. Then give them some initial data, a challenging quest, and very little time.
In our case, we were given some basic data on garnment production sites: address, number of employees, product types, retailer. Our team decided to reduce the quest to NIKE, and we immediately started our research for additional data. I always knew that there is lots of information available on the web, but still i was surprised how much we found within 10 minutes: list of all sports teams sponsored by NIKE; all NIKE shops worldwide; US and international tax reports, and much more.
We decided to compare production locations to sales activities, and visualize them on a map. Our “expedition guides” recommended a free map service – cartoDB – and within 20 minutes we had the first data on the map. The next 30 minutes we used for cleaning and combining the data, and bringing everything on the same map.
At the end, when the data expedition ended after one hour, we had two separate maps that showed our data data; we hadn’t managed to bring everything in the same map. For this reason, we decided to do it “as homework”, and completed the tasks a few days later. In fact, we invested a few more hours, added data about sports sponsoring and did some layout improvement
NIKE Activities – click to start!
My conclusion: I was really impressed how much you can achieve within one hour! Starting with almost nothing, we decided what to do, found the necessary data, and produced a first draft of our interactive map about NIKE.
Our recent paper on ”Potential and Limitations of Commercial Sentiment Detection Tools” (see this blog post) received alot of attention in the community. In face, we got several requests to provide access our data and the test corpora.
You can find our results and data at our sentiment analysis site. Unfortunaltely, we cannot provide the corpora directly, due to legal reasons. But you can find and download them from the following sources:
Hope this simplifies your work!
How good are commercial sentiment analysis tools? We recently tackeled this question in our research team, and evaluated the quality of 9 state-of-the-art commercial sentiment detection tools. We applied them to 30,000 short texts from various sources (tweets, news headlines, reviews etc.). The best tools have an accuracy of 75% for some document types (tweets), but the average accuracy over all documents is at best 60%. This means that even with the best tool, 4 out of 10 documents will be classified wrong.
Since we were convinced that there is still some “potential” for improvement, we combined all tools with a meta-classifier. It turned out that using a random forest classifier can improve accuracy by up to 9 percent points, in comparison to the best single tool.
Our results were published at ESSEM 2013. For more details, please see our paper.
While setting up my new website, i stumbled over my old “Teaching Statement”, which I wrote over ten years ago when I was PhD Student at ETH. At first, I intended to re-write it, but then I realized that it still perfectly describes my attitude and motivation. So here they are, my 50 cent on teaching:
“For me, teaching is fun. This does not mean that I do not take it seriously – on the contrary, I always prepare my exercise classes and presentations thoroughly and extensively. For as long as I can remember, I have enjoyed teaching. I find it extremely fascinating to help and guide students in their learning process, and it is one of the most satisfying experiences for me to see a student finally understand a subject, be it derivatives, quicksort, or how to smash a volleyball. This is why I chose to teach private lessons and exercise classes, and this is what makes me look forward to my future lectures.
In the following, I will describe some aspects of teaching that I consider very important, using the course “Programming in Java” as a (very successful) example. I taught this course in spring 2001 together with Alexander Below, another Ph.D. student at the Department of Computer Science of ETH Zurich. I initiated the course, driven by my first experiences in lecturing in summer 2000, and motivated by the conviction that one can learn programming only by implementing a large and complex program. The course was attended by 50 students with basic programming knowledge and consisted of two parts: The first week was dedicated to learning the basics of Java; in the remaining two weeks, the students were split into groups of five students to implement a large project, a computer game. The website of the course (in German) can be accessed via my homepage.
In the first week, we taught two lectures per day, each followed by two hours of exercises. The goal of each lecture was to introduce a “small” basic topic, such as object creation or event handling. The main purpose of these lectures was to give the students an impression of what is possible in Java, and how to use it. It was (and still is) our conviction that we cannot teach something to the students – they have to learn it by themselves. Accordingly, we only presented the subjects that we considered relevant, and left it to the students to learn the details by themselves. Our lectures were supported by exercises, where the students solved tasks at the computer. The level of the tasks was increasing from very easy to challenging, allowing each student to learn and succeed at his/her personal level. During the exercises, we were present in the computer room to answer questions; however, our main intention was to help the students help themselves. To this end, we often pointed the students only to the corresponding chapter of a book, encouraging them to find the solution by themselves, or sent them to ask another student who had solved the problem already. Especially the latter method proved to be very successful, since the students were always happy to share their know–how.
In the remaining two weeks of the course, the students implemented the game “Crazy Maze”. We had chosen this rather simple game, because we wanted to ensure that all students end up with a running program. This was in accordance with our belief that success and positive feedback is one of the main ingredients of successful learning. During the first days of the project, we helped the students every now and then, especially with the object–oriented design. Later, we reduced our interaction more and more, and the students worked autonomously. At the end of the project, our strategy had worked out well: Every group had implemented the game, and most of them had even time to add features like multi–user mode or computer players.” (mark, 2003)