Attitudes play an important role in students' academic achievement and retention, yet quality tools to measure them are not readily available in the new field of data science. Through funding from the National Science Foundation, we are developing a family of instruments that measure attitudes toward data science in the context of an introductory, college-level course. This poster will showcase preliminary results discussing pilot instruments to assess instructors' attitudes toward teaching data science and an inventory which captures classroom characteristics. These instruments, based on Expectancy-Value Theory, will enable data science education researchers to evaluate pedagogical innovations and measure instructional effectiveness relating to student attitudes. We invite instructors of data science courses to join in this discussion and to use these instruments for their own data science education research projects.

Introductory data science courses are appearing at colleges, universities, and high schools around the country and the world. What topics do we cover in these courses, and how and why are these decisions made? How do we consider the background knowledge of our students and how they hope to utilize their skills after this course (whether professionally, additional courses, or as an engaged citizen)? In addition, the course is being taught by computer scientists, statisticians, business analysts, mathematicians, journalists, etc. Each of these disciplines approaches the topics differently. What upskilling is required of instructors to prepare them to integrate material from academic disciplines in which they were not trained into the course? How much, if any, cross-disciplinary collaboration and discussion occurs or should occur in designing this course? Participants in this birds-of-a-feather will share their decision processes and choices about introductory data science courses that they teach or are designing. This includes choices made about the content as well as whether and how upskilling occurs. They will review and refine a list of current data science topics created based on national surveys of data science instructors as well as a review of curriculum guidelines. Close attention will be paid to differing language between data science instructors from different academic backgrounds. We welcome new and experienced data science instructors, educators planning on or interested in teaching such a course.

Data Science is one of the fastest growing fields with unmet demand from employers. Many academic institutions have taken on the task of creating programs to meet both current and future needs and demands. Data science, as a field, integrates aspects of computer science, statistics, and subject matter expertise which encourages cross-disciplinary conversations and collaboration. In this talk, we present results from a broad survey of instructors of introductory college-level data science courses for undergraduates. In addition, we explore the alignment of these findings with the recommendations of various professional organizations.

Attitudes play an important role in students' academic achievement and retention, yet quality tools to measure them are not readily available in the new field of data science. Through funding from the National Science Foundation, we are developing instruments that measure attitudes toward data science in the context of an introductory, college-level course. This talk focuses on preliminary work to develop the Student Survey of Motivational Attitudes toward Data Science (S-SOMADS), which includes construct development, item creation involving content experts from numerous disciplines, pilot testing, and item refinement. Along with the student instrument, we intend to develop an instructor attitude instrument as well as an environment inventory to capture instructor attitudes and course characteristics that may be associated with attitudes. This family of instruments, based on Expectancy-Value Theory, will enable data science education researchers to evaluate pedagogical innovations, create course assessments, and measure instructional effectiveness relating to student attitudes. We invite instructors of data science courses to participate in this project to discuss data science topics, serve as subject matter experts, or to collect pilot data from themselves and their students.

Data scientists - practitioners, researchers, and educators - often disagree on the definition of data science. Data science courses are taught by faculty in departments of computer science, statistics, and business analytics as well allied fields. Some data scientists create scalable dashboards, others employ and (hopefully) explain machine learning models, most (but not all) demand data wrangling and visualization skills, and some require theoretical knowledge to develop novel algorithmic and/or analytic techniques. How, then, do we determine what content should be covered in an introductory college-level data science course? Some serve as prerequisites while others are terminal experiences for non-majors (who we secretly hope to entice into the field). Some courses require prior knowledge of statistical methods, databases, or programming, uniting heterogeneous populations of students and demanding flexibility regarding programming language(s) utilized or allowed. Realistically, we often make expeditious choices because competing demands leave us hard-pressed to keep up-to-date with current skills and practices. BoF participants will share their decision processes and choices about content for an introductory data science course. Information will be shared, from participants and disseminated material, on resources and curriculum guides from various professional organizations. We welcome new and experienced data science instructors, educators planning on or interested in teaching such a course, and industrial practitioners experienced in working with or hiring undergraduates.

Data science encompasses elements of statistics, computer science, and mathematics as well as domain-specific knowledge, suggesting that with interdisciplinary conversations and collaborations we can make data science curricula more comprehensive and successful. Development of such programs provide an exciting collaboration opportunity for faculty in statistical and computer and mathematical sciences as well as a wide array of other disciplines. This BOF provides an opportunity for faculty involved/interested in the development of undergraduate data science curricula to come together and discuss forming collaborations across disciplines and what the disciplines can learn from each other. The discussion will be led by faculty from computer science and statistics departments who have been involved with both disciplinary and interdisciplinary data science education initiatives. One of the goals of this BOF is to jumpstart conversations across disciplines, which we hope the audience will continue at their home and/or nearby institutions.

Two faculty teams worked to develop significant content that could be presented in a variety of contexts to students with interests in many areas. This poster reports on the opportunities and the challenges encountered.

Participants in this Birds-of-Feather session will share thoughts and experiences in offering an introductory course in this exploding field. It will be the third consecutive year that these facilitators will conduct a SIGCSE session on a first course in Data Science. The interest in the field has increased greatly over the years, and attendees of varying backgrounds, points of view and experience are welcome. Whether the course will serve as an outreach vehicle to students of all majors, including non-technical ones, or as a first course in a formal program of some sort, we will be exchanging points of view regarding both topic coverage and pedagogical approaches. Regarding topics, we'll raise questions on the relative emphases of statistics, programming (should R be the vehicle? how about Python? other languages?), machine learning (which algorithms?), other tools, and appropriate data sets. Participants who have taught such a course will be invited to describe some of their approaches, especially if they've used such active learning methods as flipped classrooms. The facilitators, PI's on an NSF IUSE grant now winding down, bring experience from two institutions and three departments, including one interdisciplinary course. They hope to expand the community begun through the grant and these sessions, hosted at http://computingportal.org/DataScienceCommunity.

Data Science, often described at the intersection of computer science, statistical thinking and analysis, and subject matter expertise, has seen an exponential growth in the past few years. Courses (and entire programs) have been appearing at such a fast rate at most institutions of higher education, as well as some high schools, that comparisons between curricular and delivery models and rigorous discipline-based education research are often overlooked in order to gain competitive advantages. This study attempts to rectify that absence by evaluating, comparing, and discussing four different courses offered at two different institutions of higher education. Funded by NSF via a collaborative grant (DUE-1432438), faculty from Computer Science and Statistics departments collaborated on the development and evaluation of introductory courses in Data Science for all students, using a discipline-based education research approach. Data on students were gathered including demographics, curriculum, statistical knowledge, and attitudes towards Data Science. Post-course growth was measured, when available, and compared through formal statistical inference. End-of-course evaluations, with supplemental questions about student learning, were reviewed and will be summarized. Finally, reflections on successes, challenges, and lessons learned will be shared.

In this poster the authors report on experiences in teaching an introductory course in Data Science at two different institutions. Their approaches were informed by the aims of their NSF-funded project: to provide insight on learning goals, central data science topics, content modules, and a framework for implementing a flipped classroom approach to introduce data science to students with various technical backgrounds. The authors, investigators on the grant mentioned above, are a collaborative team of computer scientists and a statistician working to create flipped material for an introductory data science class. After ITiCSE the materials described in the poster will continue to be available in Ensemble, at http://computingportal.org/datascienceflipped

The use and analysis of large quantities of data have become ubiquitous in nearly every discipline. We began a discussion of the role of data science across disciplines, and the role of computing in data science programs, at SIGCSE 16. The session was well attended and the discussion was valuable. Since then, more work has been done and more people are engaged. This BOF will continue the discussion, including welcoming new voices. We will distribute copies of the report of the NSF sponsored workshop on Data Science education and discuss a new initiative to develop curriculum guidelines for data science programs. This initiative will be in its earliest stages by the time SIGCSE meets, so it will be an excellent opportunity to gather impressions about what are critical considerations for any such curriculum effort. We developed a mailing list from the SIGCSE 16 attendees and will use that list to promote the BOF. The BOF will engage SIGCSE participants who have views on the content and role of courses and programs in data science. In addition to the workshop report, we will describe results from an NSF IUSE grant to develop modules for use in many types of courses. These expect to make access to fundamentals elements of data science available as widely as possible. With these as a starting point, participants in the Birds of a Feather session will explore the emerging field of data science and its relationship to computer science education. Discussions will be hosted at http://computingportal.org/datascienceflipped

In this poster the authors report the approaches for presenting Data Science topics in Flipped Classroom mode, incorporating topics in Data Science into existing courses as well as in stand-alone courses. It provides an insight on listing of learning goals, central data science topics, content modules, and a framework for implementing a flipped classroom approach to introduce data science to students with limited technical backgrounds. The presenters are NSF-funded investigators on a collaborative team of computer scientists and statistician to create flipped material for an introductory data science class.