Data Nirvana

This post continues our City University of New York miniseries about a set of projects, collectively known as A2B, that are reducing barriers to associate’s to bachelor’s (vertical) transfer within the system of seven community colleges and 13 senior colleges. Here we look at features of the CUNY computer systems that have helped make one part of A2B, the Articulation of Credit Transfer Project (ACT) successful, and then consider some of the challenges we face as we work to extend ACT to systems beyond CUNY.

In 2019 the Heckscher Foundation for Children began funding Ithaka S+R for ACT’s development of data-informed technologies to improve transfer at CUNY. One of ACT’s products is a public website, with no required login, called Transfer Explorer (“T-Rex”), which shows credit transfer information, updated daily, and which is now approaching 100,000 unique users. T-Rex has attracted significant additional external funding from the Ascendium, Dell, ECMC, Ichigo and Petrie Foundations and is now an official resource for the university.

T-Rex’s success comes in large part from CUNY’s landscape of data nirvana: not only do we have an enormous amount of information to work with, but we also have the luxury of a single student information system (SIS) that serves all CUNY campuses, and a common degree audit system across campuses. Two examples of this nirvana landscape are: 1) every CUNY course has a unique identification number and 2) every CUNY person, whether faculty, staff, student, administrator (or some combination), has a unique identification number. For transfer students, this means we have direct access to a record of all the courses the student took at any CUNY college, as well as direct access to all the information about those courses, regardless of the CUNY college at which they were taken.

But data nirvana isn’t limited to what is essentially transcript information. The SIS also contains 1.6 million rules that tell how any course, taken at any CUNY college, will transfer to any other CUNY college. When a student transfers, all their prior coursework is automatically fetched from the SIS and mapped through these rules to the courses offered at the student’s new college.

And it gets even better: the degree-audit system at the transfer student’s new college automatically accesses the SIS to see exactly what requirements the student has completed based on whatever courses were transferred plus whatever courses the student completes in residence at the new college.

T-Rex’s progenitor’s goal was to make it easy for faculty and staff to understand and evaluate existing transfer rules. The current version, in addition to developing that feature, has a public-facing design that makes the transfer rules easy to view and understand. Making this information accessible to people outside an “IT priesthood” has had some significant positive effects.

First, all the information mentioned above relies on text typed into the system by humans. There can be typos in course titles or catalog descriptions or transfer credit rules, and making those typos visible in T-Rex has led stakeholders to clean information that was previously seen by only a limited audience.

Perhaps more significantly, those 1.6 million transfer rules can be extremely complex and thus difficult to understand. Although 99.6 percent of the transfer rules map one course to just one course at another campus, that still leaves over 7,100 rules for which a rule involves combinations of courses. Until T-Rex, only the anointed priesthood could see and understand these rules, but T-Rex lets anyone easily see how any course transfers in an understandable way.

Shining light on the existing transfer rules also helps us prioritize transfer-enhancement efforts. We have a rule for how every course transfers to every other college because of CUNY’s policy that all courses have to transfer with at least elective credit. However, that doesn’t mean that all credits transferred actually will contribute to the student’s degree progress. The most significant issue here is what we call “blanket credits”—course credits that don’t count toward any general education or program requirements, only toward the total credits needed for the degree. Sometimes this happens because a course at the sending college gives more credits than the matching course on the receiving side (e.g., a four-credit sending course might transfer as a three-credit course plus one blanket credit). But in many cases blanket-credit-only rules indicate courses that haven’t been evaluated for meaningful transferability. Currently, 91.8 percent of all our transfer rules have only a single, blanket-credit course on the receiving side. However, a large portion of those rules are never used, typically rules for specialized advanced courses, which students rarely try to transfer. Data nirvana lets us see which problematic rules are actually encountered by transfer students, and then target those rules for reconsideration.

Everything about our data nirvana has been sunshiny and bright with respect to how credits transfer. Things are a bit more hazy with respect to program requirements. But even the clouds here have silver linings.

We already have a common set of general education requirements across CUNY, with a policy that once a gen ed requirement has been satisfied at one college, it has been satisfied at all CUNY colleges regardless of how the specific courses used to satisfy that requirement might match particular courses at the destination college. So data nirvana, coupled with strong system-wide policies, helps us with transfer credit loss and general education requirement issues. The common degree audit system used at all CUNY colleges is a commercial product that uses a proprietary computer language to define program and degree requirements. We developed our own system that is able to interpret all 25,000 of CUNY’s blocks of requirements written in that language, and to extract program requirements for all undergraduate majors, minors, and concentrations at all CUNY community and baccalaureate colleges.

Having program requirements available separately from the degree-audit process means that we can produce the equivalent of the college bulletin version of all majors and minors across the university in a standard way, and we update those descriptions automatically whenever any requirement changes in the degree audit system.

Further, T-Rex now shows not only what courses are “equivalent” to each other across campuses, but also what requirements a course taken at any sending college can satisfy at any receiving college.

We are currently extending this work to show potential transfer students how many credits they will still need for a particular major at the receiving college. This same analysis will also allow us to show a community college student the optimal transfer point: Sometimes it’s better not to spend time/credits completing an associate’s degree if doing so doesn’t reduce the number of courses/credits that will be needed to complete a bachelor’s.

Despite data nirvana, we have had to deal with issues of data integrity and data complexity.

The data integrity issue is common to any collection of data: there were errors when converting data from previous data systems to the current one, and people make errors when entering new information. Fortunately, in our case, most of these errors have involved data items that haven’t been detected previously simply because they have never caused problems for students.

Data complexity has been much more of a challenge. We’ve been fortunate to have the support of the University Registrar, who has answered many questions we’ve had about the data structures involved, and also fortunate to have a team member who has held positions in admissions, advising, and data management across CUNY. But for others on the development team, it’s been a steep learning curve!

A particularly challenging form of complexity came as we developed the system for showing academic programs and their course requirements. The problem is that people can code similar requirements in different ways, and even code requirements in ways that don’t make sense but haven’t actually caused problems (yet!). With each of the 20 undergraduate colleges maintaining its own degree audit system instance, there’s a lot of “campus character” that goes into how requirements are coded across the University.

Our Transfer Explorer (T-Rex) has been extremely popular at CUNY, and our success has garnered national attention. We now have funding to expand T-Rex to institutions outside CUNY’s data nirvana. The key to accomplishing this is to establish common data formats for courses, transfer rules, and program requirements (and mechanisms for translating other systems into those common data formats). Other organizations, including commercial ones, have already addressed some of these issues. Our goal is to develop a platform that ties all three together, and to make the platform available to the academic community on a nonprofit basis. We’re very glad we can help CUNY students understand and explore their transfer options effectively within our system, and are looking forward to helping transfer students nationwide do the same.

Christopher Vickery is professor emeritus of computer science at Queens College, CUNY, a founding member of the ACT team, and creator of T-Rex’s progenitor.