Coursedog

Submit a Ticket My Tickets
Welcome
Login  Sign up

Understanding How Projections Work

Table of Contents 
Overview
When Are Different Algorithms Used?
Pre-Enrollment Window
Enrollment Window
Section Insights
Related Articles


Overview

  • Course Demand Projections (also known as “CDP” or “Scheduling Analytics”) should be leveraged throughout the scheduling process, from the time the schedule is rolled, through the close of the enrollment window, in order to align the schedule most closely to student need.

  • CDP uses two different algorithms, Pre-Enrollment Window and Enrollment Window, in order to predict future enrollment.

  • The algorithm that is used at any given time is dependent on where your institution is in their scheduling and registration cycle.

  • In addition to predicted demand, CDP also surfaces Section Insights that illuminate underlying drivers of student demand. 

  • Configure all settings within Academic Operations Analytics > Settings > “Analytics Configuration” to generate these insights.


When Are Different Algorithms Used?

  • CDP surfaces different insights based on the set scheduling and enrollment window timeline for the given term.

  • This timeline is set by administrators at your institution and should be visible at the top of “Courses” tab for the given term.

  • Different algorithms are used at different points in the scheduling and enrollment process in order to make the most accurate predictions based on the available data.


  • The timeline is entirely driven by the start and end dates of the enrollment window for the given term”

  • Prior to the start date of the enrollment window, CDP is in the “Pre-Enrollment Window”.

    • During this time, the “Pre-Enrollment Window” algorithm is used.

  • From the start date of the enrollment window through the end date of the enrollment window, CDP is in the “Enrollment Window”.

    • During this time the “Enrollment Window” algorithm is used.

    • The only exception to this is if there is not enough data to fuel the enrollment window algorithm; in which case, the system will fall back on the pre-enrollment window algorithm. This can occur if we do not have actual enrollment trend data for the course for the previous “like term’s” registration window.

  • Following the end date of the enrollment window, CDP is in the “Post-Enrollment Window”.

    • During this time no algorithm is used. Actual demand is simply reflected as it took place based on the actual enrollment and enrollment capacity data in the system.

    • The column will be re-labeled “‘Actual Demand” to reflect that the insight shared is no longer predictive but rather a reflection of what demand ended up being.


Pre-Enrollment Window

  • The pre-enrollment window algorithm is primarily utilized before the enrollment window opens to determine whether a course is likely to be overfilled, balanced, or underfilled.

  • These insights should be referenced as soon as the schedule is rolled in order to initially align the section offerings with student need. 

  • The projection relies on multi-variable linear regression to predict enrollment trends, the variables being: past fill rate, actual enrollment, and enrollment capacity.

  • The linear regression will output a predicted number of enrolled students. This is compared to the current aggregated enrollment capacity across the sections of the course to determine a likely fill rate. 

  • Whether or not the course is predicted to be “Likely to be overfilled, balanced, or underfilled” depends on the fill rate definitions set for each category within “analytics configuration”.

  • There are a few hard requirements in order to generate a pre-enrollment window recommendation for a course.

    • Coursedog must have at least 2 years of like-term data for the course, within the past 8 years, in order to generate a recommendation.

    • For a section of the course to be included in the regression, it must be considered “offered” based on the “section status” configuration within “analytics configuration”. Additionally, both actual enrollment and enrollment capacity on the section must be greater than 0.

  • If the above conditions are not met, the system will simply surface “N/A”.

  • It can be helpful to look at the “Enrollment” graph within a course to understand the underlying enrollment trends driving the recommendation.


Enrollment Window

  • The enrollment window algorithm is meant to support last minute adjustments to section and seat offerings to align the schedule with demand. It similarly predicts whether the given course is likely to be underfilled, balanced, or overfilled. 

  • This algorithm, as its name suggests, is used only when the enrollment window is open in the current scheduling term.

  • If recommendations were implemented during the pre-enrollment projections window, and there are no significant changes to student behavior, then fewer changes should need to be made.

  • The projection compares the current enrollment to the enrollment fill rate at the same point in time in the previous like term’s enrollment window, in relation to its final fill rate.

    • For example, if we are halfway through the current enrollment window, and the course was 90% full at this point in the last enrollment window, Coursedog will predict that the current enrollment is 90% of what the final enrollment will likely be.

    • Coursedog will then output a predicted number of students based off of this, and compare it to the current aggregated enrollment capacity across the sections of the course to determine a likely fill rate.

    • Whether or not the course is predicted to be “Likely to be overfilled, balanced, or underfilled” depends on the fill rate definitions set for each category within “Analytics Configuration”.

  • There are a few hard requirements in order to generate an enrollment window recommendation for a course.

    • The enrollment window for the last “like term” and the current term must have been set within “Analytics Configuration”.

    • The course must have been offered in the last “like term” in order to generate a recommendation.

    • Actual enrollment numbers for the course in the last “like term” must have been pulled into Coursedog regularly during the enrollment window. 

    • For a section of the course to be included, it must be considered “offered” based on the “section status” configuration within “Analytics Configuration”. Additionally, both actual enrollment and enrollment capacity on the section must be greater than 0.

  • If the above are not satisfied, the insight will 100% be based on the insight from the pre-enrollment algorithm, if present.


Section Insights

  • Section insights identify key trends in section demand based on campus and modality.

  • Coursedog will identify if there are certain campuses or modalities associated with the sections of a given course that were significantly overfilled or underfilled.

    • The threshold to determine significance is set within “Analytics configuration”.

    • E.g. If the threshold is 90%, the system will flag on a course-by-course basis whether there were any campuses or modalities that were overfilled or underfilled 90% of the time or more based on the sections offered.

  • If your institution does not use “Campus” or “Modality”, these insights will not be surfaced.


Related Article

Managing Analytics Configuration



Did you find it helpful? Yes No

Send feedback
Sorry we couldn't be helpful. Help us improve this article with your feedback.