Best practices for undergraduate statistics courses strongly emphasize analysis of real data, effective use of statistical computing software, and meaningful problem-solving and decision-making. Practitioners and popular media are also paying increasing attention to the ethical foundations and social consequences of (big) data analysis. Combining real data with community service in academic service-learning projects is a way to bring these ideas together. Calvin University recently introduced a course in Advanced Data Analysis covering modern regression techniques. As a term project for the course, student groups act as statistical consultants for non-profit organizations or academic researchers. These statistical consultancy service-learning projects allow students to put their academic and technical skills to use for community good, learning best practices and real-world skills in the process. Here, I consider case studies from the Calvin course, comparing with similar projects at other institutions. Using results from surveys of students and community partners, I will explore what makes such projects successful and valuable for all participants, as well as potential pitfalls.