Campus Analytics

Campus Analytics is an initiative of the Center for Digital Education (CEDE), with the goal of bringing together scientific and institutional research at EPFL.

The EPFL leadership is committed to transform our university into a world leader in the domain of data science. Indeed, a considerable number of research labs at the university rely on big data and machine learning techniques to conduct ever more innovative research, in domains ranging from neuroscience to social media mining.

Yet, such an ambitious goal can not be complete without the application of data science principles to the functioning of our own university, our teachers, and our students. The amount of educational data that is available today, especially since the advent of MOOCs, and with the strong online presence of the current students’ generation, has reached an order of magnitude that allows it to be called “big data”. And where “big data” exists, so does machine learning and data visualisation (not just excel charts, but also scientific visualisation).

Applying data science principles to EPFL academic data can unlock secrets and provide insights that would otherwise be concealed from normal observation. It can show how students use different learning strategies to reach a common goal, how a specific lecture is causing students to dropout of a course, how a specific choice of courses can lead to better or worse results in future courses, how learning paths lead to different careers, and how well the general focus of the EPFL curriculum is adapted to the current reality of the academic and business worlds.

These are all questions that affect decision-making at EPFL, whether it’s decisions taken by students, professors, or policy makers. Our goal is to have at least some of these decisions be informed by data, by leveraging the research and brainpower in our research labs.

Example Projects

There are a variety of projects related to Campus Analytics, that are in the horizon. Here’s a few examples:

– Building a course recommendation system for students based on machine learning;

– Estimating course dependencies in the EPFL curriculum, based on course description analysis and student performance (pre-requisites analysis);

– Estimating the risk of a student dropping out of a course, in order to create an early alert system;

– Evaluating learning strategies students use at each point in a course, and determine which ones lead faster to success or failure;

– Studying if remediation activities and additional resources provided by the teachers are effective ways to enhance students’ understanding;

– Providing students with better (data based) career planning recommendations, such as specific course choices that lead to a certain desired career;

– Providing students with recommendations for startup companies, and which study paths lead to areas of high venture capital investment.

Data Privacy

The Campus Analytics database is stored and managed by the CEDE, under the same security standards that apply to IS Academia. No personally-identifiable information (PII) related to our students is to be disclosed or shared with people outside the CEDE. When data subsets are shared with research labs at EPFL, students are no more than anonymous data points identified by irreversible hashed IDs. All research projects are bounded by a non-disclosure agreement (NDA) signed by lab directors.

For inquires please contact Patrick.Jermann@epfl.ch.