April 2, 2019 | Institutional Research, Predictive Analytics

    NCAIR 2019: “Treasure” Found in Every Presentation!

    Last week, representatives from more than 40 institutions across North Carolina convened at Wrightsville Beach for the annual NCAIR conference. Located in a beach town near waters once frequented by the infamous privateer Blackbeard, this year’s event was piratically titled “Finding Your Treasure." Directors, researchers, and analysts discussed how they discover hidden gems and insights among the sometimes-overwhelming amounts of institutional data.

    A common refrain among presenters and attendees was the importance of clearly understanding the research question. UNC-Chapel Hill’s Quin Jernigan emphasized this in her session on data collection with web-based survey tools, where she engaged the audience in a lively presentation and discussion of the pitfalls of surveys whose purpose and design have not been fully thought out. Elizabeth Reilly and Callie Uffman from the UNC System Office shared an overview of the UNC System Strategic Plan metrics and annual indicators. These have been carefully defined according to the goals and focus of each UNC System institution, and visualized in interactive dashboards available to the public in order to intuitively show progress towards the goal.

    The use of predictive techniques was also a recurring theme. Giana Malak of the UNC System Office shared how her team used data from all UNC system institutions to create predictive models of graduation and retention solely from student application data – and then were able to successfully translate these findings into actionable insights for the Board of Governors. Mondrail Myrick from UNC - Pembroke reviewed his findings from a logistic regression on student retention as it relates to a student’s out-of-pocket tuition cost. The focus of both sessions was on the inputs to the model and the extent to which those factors affect the variable of interest, rather than just the predictive power of the model.

    For institutions curious about developing their own predictive models to answer questions and optimize resources, I presented a practical, no-nonsense guide on how to construct a successful data science project from conception to implementation. At HelioCampus, I apply predictive techniques to institutions’ data to help them optimize their resources and achieve their goals for student and institutional success. If you missed the session or are interested in learning how HelioCampus can use data science to unlock the treasures in your own data, let us know!

    Related Posts