In this webinar, we’ll walk through ways to unify data systems and sources to build a foundation for improving student outcomes.
IT organizations at many higher education institutions are getting stretched beyond their capacity. They are realizing that their existing technology investments may not be able to live up to the expectations of institutional teams that need technology to make more data-informed decisions to improve student success. Some institutions’ data are stored in systems that are obfuscated, 10 or more years old, with poor documentation, disorganized, inaccessible, and disparate. Other institutions where data may be more easily available are facing the fact that its quality is erratic across many dimensions. And still other institutions with legacy LMS and SIS systems may find that these have become shadow systems due to developers leaving, and the institution itself evolving over time.
Getting to a Holistic View of Data
In this webinar, Civitas Learning vice president of Engineering, Tom Warmbrodt, and Senior Director of Client Success, Dr. Matthew Milliron, discussed the challenge of closed systems, semi-closed systems, and impenetrable data that make it difficult for institutions to view students in a holistic way. Attendees learned how disparate systems and data sources can be brought together to create a single student record, to enable institutions to make educated comparisons and assess what initiatives are having an impact on students.
The Challenge: Disparate Data Systems
“There are numerous disparate systems scattered throughout the institutional IT ecosystem, or as I like to call it, the academic graph,” said website moderator, Tom Warmbrodt. “These include your traditional student information systems (SIS) and systems that may provide insight into engagement and learning behaviors, like your learning management systems (LMS), grade books, facilities usage, advising data and more.” He says some systems are well understood and supported and others are shadow systems that collect data but rarely, if ever, does that data get analyzed in any way. On top of that there are all sorts of systems that get used organically by students and faculty: social networks like Facebook, instructional data from YouTube. “So the challenge is how can we know what is working and what’s not, when and for whom? At Civitas Learning, we unify data. We’ve had to solve this problem for our partners.”
“This image illustrates an approximation of some of the potential data sources,” said Warmbrodt. “Bringing all of these data streams together is called data fusion. The four-part process involves ingesting the data, mapping the data, performing normalization, and finally, performing data availability segmentation. For us the goal is to generate actionable insights from predictions. We do this to predict successful course completion, persistence term-to-term and graduation.”
Comprehensive Data Drives Better Insights
When he joined Civitas Learning three years ago as the technical co-founder, Warmbrodt hypothesized that all colleges and universities would need the same 50 – 100 data points to drive predictive insights, and that once Civitas was working with a handful of institutional partners he would see congruity and be able to just select that specific data from each institution. “We’re at more than 40 institutions now,” said Warmbrodt, “and I’m surprised how wrong that hypothesis was. If you created a Venn diagram of all of the most predictive data points across all of our partner institutions,
there would be some common ground, but much more interesting is how every college and university has something special, something unique that creates insights.
Institutions are Unique
“It makes sense if you think about it,” he explains. “Even if two different institutions have the same SIS and LMS, how they use them is different. Their student populations are different. The initiatives they’ve tried are different. So you really have to go broad AND deep to drive great insight. We call this process data discovery. We find the potential data sources, including the shadow data sources – that’s the broad part. Getting into the data and understanding what it means, how we might stage it, map it, canonicalize it, and then create and compete features — that’s going deep. Feature creation is key in getting meaningful actionable insights. If you do both, you enable a lot of insight; a lot of predictive power, and you can segment and cluster the data in some very informative ways. This helps make precise decisions that are targeted to specific students.”
View the Webinar to Learn More
Register with the form above for immediate access to the archived webinar to learn more about the benefits of feature creation, extraction and competition and importance of data availability segmentation.
Dr. Matthew Milliron, Senior Director, Client Success
Over the last 15 years Dr. Millliron has led technology programs and organizations in both education and non-profit associations. Milliron most recently served as the Chief Information Officer for EDUCAUSE, and in addition to his leadership roles, he has taught courses on eLearning design and presented on technology and higher education topics.
Tom Warmbrodt, former Vice President of Engineering
Warmbrodt oversaw Civitas Learning’s development and information technology teams while working closely with partners to guarantee seamless and secure integrations. Prior to that, he served higher education and K-12 for a decade as the chief technologist for Digital River Education Services, JourneyEd.com, and Academic Superstore.