Visit Modern Campus

From Reports to Analysis: Big Data and Data Analytics in Small Institutions

The EvoLLLution | From Reports to Analysis: Big Data and Data Analytics in Small Institutions
The adoption of data analytics and Big Data in higher education has led to an improved staff experience, greater operational efficiency and effectiveness, and the ability to improve the student experience.

Big Data, a highly subjective term, entered the Gartner Hype Cycle in 2013, and rather than moving through the various stages of that cycle, it fell off the cycle in 2015. This surprising move was explained by the title of a blog post by Gartner’s Nick Heudecker: “Big Data Isn’t Obsolete. It’s Normal.”

The reality is that Big Data has been the norm in many areas of science for a long time, but now it has become the norm everywhere! The big question is, has Big Data become the norm in small higher ed institutions?

Data analytics is widely defined as “the discovery, interpretation, and communication of meaningful patterns in data.” No one can argue about the value of this until you drill deep down into it. The interpretation of data is a particularly dangerous business and it is tightly coupled to the other landmine: communicating the discovery. It is fair to say that the interpretation and communication are much bigger issues in smaller institutions than larger ones.

First off, Big Data itself is relative and perhaps it is time that we seek a definition, even if it is loose. Lacking that, traditionally, we have considered the data stored in ERP systems as institutional data. Increasingly, with the proliferation of ancillary systems that we use to support many of the administrative functions of the institution, data warehouses that attempt to integrate most of those data have shifted to become the institutional data source.

However, there are terabytes of other data to which we have access that we do not track or attempt to store—mostly for privacy reasons. This includes data such as LMS activity and course-related multimedia usage, where practically every click can be tracked and stored for further analysis. Similarly, we can discern a lot based on email logs or network usage, which many institutions do not want to engage in because of privacy. There’s a wide and diverse range of technologies available for storing and analyzing these data—the roadblock is not availability. Instead, other factors such as the campus culture, and resources (both in terms of personnel and financial) necessary for supporting and analyzing the information both play a crucial role in developing and executing a strategy.

Starting this conversation with endpoints like data analytics and Big Data is a bad strategy, because this always results in solutions in search of questions. On the other hand, many of us suffer from a common set of problems—lack of institutional data definitions, data governance, old-fashioned reporting (rather than analytics) and de-fragmentation of institutional data—for which data analytics and techniques for handling Big Data can provide solutions. It is also important to highlight the advantages of data analytics over traditional reporting when we have conversations with our users. Reporting suffers from the fact that it is highly targeted and the requestor needs to have a deep understanding of what data they need to answer a particular set of questions. Analytics is more for exploration that allows interesting information—that would otherwise take multiple report requests to satisfy—to be extracted and shared. Each have their own clear advantages and disadvantages, and over a period of time, a seasoned professional can learn to use both reporting and analytics appropriately.

At Wellesley, we have been successful in executing a strategy that began approximately five years ago primarily to address the data needs of the community. Rather than being reactive to find a short-term solution, we began with a formal governance committee of highly committed individuals with support from the president’s senior leadership. This is so important for the success of a project of this level of importance.

We took almost a year to agree on institutional data definitions through an iterative process, and after almost a year of delay because of personnel changes, we rolled out data analytics systems for institutional finance, and student and financial aid data. The student system is the most widely used and appreciated while adoption of the finance and financial aid systems have been low.

The student analytics system, called WANDA, has been in use for three years. One of the most often used features of WANDA is an exploratory dashboard for academic departments with ten years of enrollment data in a variety of tabular and graphic forms. This allows a faculty or staff member to easily explore data about course enrollments and additional dimensions such as majors/minors. Departments use data from WANDA to draft reports for external review committees, make informed decisions about which courses to offer or do capacity planning based on trends that they see. The fact that this is a self-service analytics platform with institutional definitions has reduced certain types of data requests directed to offices such as Institutional Research and the registrar. In collaboration with these offices, we have been able to enhance WANDA based on the new requests that are directed to them.

The fact that a faculty member in the humanities, in collaboration with one of our staff, was able to probe WANDA to understand some of the drivers for the decline in enrollments in humanities courses is an example of how a reliable and easy-to-use data analytics system can be beneficial to a small institution. Each example that we come up with in the data analytics world could have been done using the reporting paradigm. The difference is the time and effort it would have taken to do the same thing. This is a good enough reason to do it. Unfortunately, when a new solution like data analytics is an abstraction, it is extremely hard to convince anyone that it is far superior to status quo.

As a small residential liberal arts college, data analytics concerning student success and retention are far more complex than at other institutions. Our current model of strong relationships between faculty and students, as well as between staff and students, provides us a view into these subjects that’s much different than what some of the larger institutions have. As a result, we are being very careful and ask the fundamental question, “What are some of our challenges with respect to student success and retention?” and look for ways to address them using WANDA. Similarly, there are many other areas that we will soon be looking at—class and classroom utilization to name a couple.

Emboldened by the success of WANDA, we recently launched WENDY, a data analytics system for alumnae data. The WENDY dashboard allows us to explore alumnae demographic data including information such as occupations, alum clubs and interest groups. This is perfect for an office where the staff spend literally hours to generate reports that show very specific slices of data. Now, the Alumnae Office can produce a report that creates a list of living alumnae in California who are lawyers who graduated during the ‘90s who want to receive communications from the College in seconds.

Based on our MOOC experiences through edX, we value and appreciate the possibilities of data analytics and Big Data in improving teaching and learning. In this case, we have data about learners at such a granular level that we can begin to understand learning styles and help structure the learning process accordingly. Increasing experimentation of blended learning in the liberal arts combined with the importance of assessment will open up possibilities for the use of Big Data methodologies and data analytics in our space.

I believe that we are now in a good position to collaborate with others at the college to help because we have managed to do some of the first steps correctly.

Author Perspective: