Abstract:

When choosing mentors for research projects, seminars or theses, students are usually inclined towards professors who they personally met or have experience with from past courses. In larger universities, when choosing a mentor, the students rarely have experience with all the professors or are aware of who else works in their general area of interest. Professors might teach in other courses or teach to different student groups, and non-lecturing researchers are rarely involved in the teaching process. Students sometimes even turn to supposed word-of-mouth advice of their peers, usually over social networks. This creates a risk of missed opportunity, as there might be a mentor that is better aligned to the interests of each student.

To help students make better and more informed choices, we have devised an automated approach to allocation of the true students and mentors interests to a body of knowledge map, which then enables visual exploration of the domain by students and mentors. The system works in two phases, a preparatory phase and exploration phase.

As part of the preparatory phase, referent map of the body of knowledge of the entire field of interest is established - a multi-dimensional topic analysis model including the mapping of all relevant knowledge areas, units and topics, including their in-between relations. Then all the discovered and published research papers from each prospective mentor are analyzed and pin-pointed to the body of knowledge map and to specific point-in-time. The map is regularly updated with new publications as they become available in the institutional research work repository. In addition, the prospective mentors can upload other documents to the system that further expand the list of their competencies. With this process, the referent map that is created showcases the entire knowledge domain "geography" and each research work is laid down in the most appropriate location in that map.

In the exploratory phase, students explore the map in a visual process. The map shows the domain "geography" of all knowledge areas, units and topics, and points of interest are added to specific locations in the map that represent each specific research work that is part of the database. The students can navigate the map and explore the topics they are interested in and discover who else works in that topic. In addition, students can upload paragraphs of text where they express specific interests and the system finds the nearest location in the map related to that interest. The students can also highlight their domains of confidence -- map locations based on descriptions of projects they have successfully completed and courses they have passed with higher grades. Then they can visually distinguish who else works in the same domain, and discover prospective mentors who truly have the most similar interests. The map can also be used to show how interests have changed over time and help the students to discover if interests migrate from one area to another, so they better align what they work on now, to current trends.

The system is offered to students as an optional tool, and the use is not mandatory. Preliminary explorations of the system by students indicate positive experience and satisfaction. While the system was tested in the computer-science domain, it can translate to any other domain. The system is published as open-sourced and members of the community are invited for future collaboration.

Keywords:

Subject field, body of knowledge, visual mapping, mentor discovery, competence mapping, research interest mapping.