BLOG

The Data Studio: Innovating How We Learn and Teach Data Science and AI

16 April 2020
- by Nanda Piersma, Lector Urban Analytics, Scientific Director HvA Expertise Centre Applied AI, Founder of the Data Studio and Researcher at CWI

There is an international shortage of Data Science students. Educators need to scale up and update their programmes in an attempt to meet the increasing demand for well-educated, talented students. Is this “yet another hype” or do we need to innovate education in these fields in a structural way?

The Data Studio, funded by a KNAW Comenius Senior Fellow grant, at the Amsterdam University of Applied Science is innovating education programmes for data scientists. We see this innovation as a key element in meeting the needs of students in the fields of Data Science and AI.

The Data Studio

Co-creation is the foundation of our philosophy. The Data Studio provides a safe environment in which everyone, be they student, lecturer, researcher or practitioner, has the intrinsic motivation to learn all the time. In the fast-moving field of Data Science, we are all students.

Data Science fundamentals include the Data Science community. We provide students with foundational theory and programming skills. This is taught together with data ethics, project management, communication and business skills in the context of a real-world data science project. Just as in a “real job”, students have access to a network of professional data scientists, open source code, meetups and the community nurtured by Amsterdam Data Science. Our main goal is to allow students to experience the passion we all have for the field.

Education programmes in the studio are practice-based research. When projects start we do not know the answer to the research question. We break down the research into projects suited to the different levels of students, blurring the distinction between teaching and research and between lecturer and student.

Projects are organized as a project chain. Students who finish a project transfer the results, the methodology and the code to the researchers and the next group of students.   This introduces challenges in project continuity and educationally appropriate match-making of students and projects. 

Teething Problems and Success Stories

Starting in January 2018, we have experience of four semesters of dedicated Data Science Bachelor programmes in Aviation, Applied Mathematics, HBO ICT and Communication & Multi-media design, as well as research programmes, in sustainable energy systems and urban analytics. 600 students in 45 projects have participated and we believe we can scale up.

There were teething problems, of course: projects lacked data, had unclear research goals, lacked student involvement, or were aimed at an inappropriate student skill level. 

There were also very successful projects, such as on waste collection in cities. Using open datasets of complaints about public space (MORA) and waste collection from the City of Amsterdam, we created a dashboard of underground containers, a mathematical model for the prediction of future waste volumes, a model for the optimal location of new underground containers in IJburg, and a dashboard matching complaints and compliments data for street-waste neighborhoods.

The Importance of Data and Collaboration in Uncertain Times

At the time of writing, all universities and schools are closed because of the COVID-19 crisis. Our challenge now is to educate without real-life contact, and missing out on the fun of learning and programming together. Crises stimulate creativity and inventivity, resulting in transferring the learning programme to virtual classrooms, video clips and online meetings in three days. 

To continue our co-creation philosophy in our new online existence we implemented the following:

  1. Online hackathon: using the virtual classroom instruments available we created a four-day online hackathon only nine days after the HvA was closed. 
  2. Co-creating video clips: We asked our students to make a video clip (in Dutch) explaining a Machine Learning technique using an open data set, complemented by a tech report. The results are published on www.datastudio.amsterdam. This assignment replaced the exam for the data science fundamental period, at the same time providing more online educational material than can be achieved in a traditional educational setting.
  3. Data Science for real: We invited students to join the HackCorona community call from big data natives to help fight the corona crisis.

As the prime minister and the King of the Netherlands told us in their speeches in March: we are separated, but we need to continue our social commitment in creative ways. We hope that we will continue to provide the Data Science community with well-educated talent, with the ethical and moral know-how to do the right thing, from the heart and with a passion for Data Science.

Comments are closed.