The project has two main objectives: i) Developing a semi-automatic methodology for curriculum design based on text mining and a rich set of information ii) Designing a reference curricula for data science and BPM based on the proposed methodology Traditionally, curriculum design is a tedious manual endeavor performed by costly experts. Decisions about what content to include in a curriculum and what competences to teach are made that are based on highly subjective impressions of individuals. These decisions are very rarely backed with solid quantitative measures. Furthermore, since curricula are designed by academics there is a high risk towards content that might be primarily of academic concern. Given the abundance of information that is available such as job ads (reflecting industry needs), curricula from other universities (reflecting the state-of-the-art) as well as a myriad of scientifically published articles and books (reflecting future trends), it seems necessarily to exploit this information for curriculum design. European initiatives clearly stress this point, e.g. the Digital Agenda 2020 stresses the importance to create more and better standards in Europe. Our work might be an important and disruptive step in improving decades (if not centuries) of manual, subjective curriculum design. To this end, we rely on information found on the internet as well as sophisticated text mining techniques that are applied (and, potentially, tailored) to the problem of curricula design. We believe that our new process does not only improve quality of curricula as well as transparency, an important value in Europe’s society, but also reduces development time in the long run. This could free up valuable resources and make them focus more on additional tasks that create value for Europe in business and society. We also believe that such a process can only be created by a consortia of multiple universities with highly qualified people bringing different views on curriculum creation. The field of information systems is quickly evolving. Therefore, there is a constant demand to update curricula to reflect new trends in research and industry. This project will also have as output two curricula for two (sub)disciplines, namely, data science and BPM. Data science spans several well-established disciplines covering statistics (mathematics), visualization (computer science) and machine learning (computer science). In particular, data science, which is a rather new discipline, undergoes many changes with more specializations evolving. We believe that a consortium of multiple universities is better at validating the curricula. Furthermore, we hope that this project will further deepen cooperation to offer graduate programmes with adequate specializations that best fulfil the needs of Europe’s economy.
Context / background of the project: 'Text Mining for Curriculum Design for Multiple Information Systems Disciplines' is a cross-country strategic partnership focusing on the challenge of designing a data-driven curriculum in a European context. Curriculum design today heavily relies on manual efforts of experienced academics. While they are undoubtedly competent individuals in their field, this process is highly subjective and difficult to be objectively evaluated. Furthermore, since the design is typically primarily done by academics only, important perspective from industries might be neglected. This project contributes to improving the process of curriculum design by developing a data-driven methodology, using text mining, for curriculum design in order to incorporate information from other stakeholders such as industry and students. By combining the expertise from three partner institutions in Liechtenstein, Germany, and Ireland, a methodology for curriculum design using text mining is developed and evaluated. Furthermore, an open source software supporting the methodology is developed, tested, evaluated and made available to the public. Objectives: 1) Developing a semi-automatic methodology for curriculum design based on text mining and a rich set of information; 2) Designing a reference curriculum for data science extended to related, upcoming topics such as artificial intelligence and advanced machine learning based on the proposed methodology. Number and profile of participating organizations: • University Liechtenstein (Lead Partner): an internationally recognized top-rated university, with teaching and research being closely interlinked. • Westfälische Wilhelms-Universität Münster (WWU), hosts the Münster School of Business and Economics (MSBE), one of the largest and most prestigious business schools in Germany evidenced by the top ratings consistently achieved in all university rankings. • National University of Ireland Galway (NUI Galway), the leading university in West of Ireland and internationally top ranked university, offering a wide range of both under- and post- graduate courses across five colleges, 16 schools and over 60 academic disciplines. Description of undertaken main activities: The intellectual outputs of this project were: • 01: Methodology for curriculuum design using text mining. A process description of the data driven curriculum design. Available as a webtool at https://mycdp.eu/. • 02: Open source software supporting the methodology. A text mining tool to support the proposed curriculum design process. Available as an open source code repository at https://github.com/joshandali52/textis • 03: Case studies: Curriculum design for Data Science and related, upcoming topics such as artificial ingelligence. Assesment of Data Science curricula of participating organizations (see Impact for details), which processes were resulted in research publications (see Description of the Project for details) • 04: Final Report. Final report of the project, providing an overview of the project, including descriptions of IOs and follow-up plans. Project results and impact attained: Curriculum design is a key concern for universities across Europe as it decides what content to include in a curriculum and what competences to teach. A semi-automatic data driven curriculum design process is aimed to outputs curricula which: i) incorporate views from multiple stakeholders (eg. companies, students), ii) allow quantitative assessment, and iii) can be designed with reduced manual work. The proposed curriculum design process has been made publicly available, supported by an open source software, and evaluated through the development of data science curricula in participating universities. • A semi-automatic data driven curriculum design process, leveraging various data sources such as job advertisements and student feedback • An open source software as supporting tool, integrating text mining techniques to process and analyze the available data sources • Evaluation and implementation of proposed curriculum design process • Dissemination activities such as conference contributions, workshops and publications The project has led to the development of curricula for the Data Science specialization in University of Liechtenstein and assessment of the Data Science curriculum at NUI, Galway. Both were parts of larger curriculum updates in both institutions. These updated curricula are expected to improve the relevenace of lecture contents, provide better service to the students, and thus preparing better graduates for the larger society. Exchanges on the project have also contributed to an Erasmus+ project coordinated by University of Liechtenstein on “Meeting Industrial Demand for Skills in Information Security Education”, continuing the effort of improving higher education curriculum, in this case within the discipline of Information Security.