As directors, faculty, and program leads, much of our work around data science for social impact rests in various administrative tasks. Importantly, we ensure that through these tasks we are consistently providing opportunities for students to frame and define social impact, relate these considerations to already planned research and partnerships, and/or utilize these engagements to consider new directions for research practice. In doing so, the frameworks and models developed can be truly student-centered and engaging. Social impact can be defined in a multitude of ways, using various methodologies, and one way to help students consider their work and impact is to begin by thinking through impact within the local contexts.

The consideration of social impact can relate to many domains of practice, and we saw evidence of these various domains across our institutions and our practices. There is no single definition of social impact but there are differences in how we might frame, for example, equality, equity, and justice. Similarly, the work of social impact requires a more complex framing in many instances that considers the historical, contemporary, and future goals of any given organization. Considering the role of your organization and the target audience is often best done in collaboration with local organizations and partnerships, or long-standing relationships that can be leveraged by leaders.

The lesson here was multifold.  It sometimes takes time for a new course to catch student attention and get uptake through word of mouth.  New courses taught as special topics are not always “counted” in the same way towards credentials and degrees, so, for example, they might not show up on a list of courses that satisfy graduation requirements.  It was also tricky with a new course sometimes to figure out which level to offer it at.  The course number can be a signal about how advanced the course is and who is welcome in the course.

For example, the Data Science Academy at NC State didn’t have an internship program, but we were able to build data components into existing paid internship programs that were serving groups aligned with data.org’s mission.

Even when we didn’t have specific advice for each other, it was encouraging to know that many different types of institutions had similar challenges when it came to starting up new programs and navigating the interdisciplinarity of data science.