Student engagement and analytics systems and methods with machine learning student behaviors based on objective measures of student engagement
Embodiments leverage wireless data passively collected by access points of a school to correlate a student's behavior with respect to time and place. This data-driven approach can quantify student behaviors and objectively measure and track how students move and interact on campus. For example, if a student accesses a wireless router that is in the same physical location as a class on the student's schedule at the time the class takes place, the correlated time and place quantifies how the student behaves with respect to class attendance. The invention takes the sum of such interactions (e.g., attending classes, studying in the library, etc.) and produces a student engagement score (SES) for each student. The SES is evaluated to determine how student behaviors change over time (e.g., throughout a semester) and whether the student is trending to a low engagement score and thus “at-risk” of dropping out.
1. A method, comprising:
processing wireless local area networking data into data sets having an internal data structure, the wireless local area networking data passively collected by wireless access points of a school, the processing performed by a server computer operating on a student engagement and analytics platform;
analyzing the data sets to produce objective measures of student behaviors specific to a student of the school, the analyzing performed by the server computer and comprising:
correlating place and time information in the data sets specific to the student;
associating the place and time information to a student identifier of the student and to a longitude and latitude coordinate of a physical location at the school; and
generating a plurality of behavior tables, each behavior table of the plurality of behavior tables representing an objective measure of a student behavior;
performing, by the server computer, machine learning of student behaviors utilizing the plurality of behavior tables, the performing comprising:
constructing a multivariate linear regression model utilizing a set of student behavior variables;
solving the multivariate linear regression model to obtain an actual student engagement score representing an overall student retention of the student; and
generating a forecast of future behaviors of the student utilizing the actual student engagement score;
performing, by the server computer, at-risk student engagement detection utilizing the actual student engagement score, the at-risk student engagement detection comprising:
determining at-risk student retention thresholds using machine learning; and
determining an at-risk student engagement detection outcome utilizing the at-risk student retention thresholds; and
updating a user interface on a client device in real time to include the at-risk student engagement detection outcome.
2. The method according to claim 1 , wherein the internal data structure is internal to the student engagement and analytics platform and has a set of attributes, the set of attributes including at least a source network address, a destination network address, and a timestamp.
3. The method according to claim 1 , wherein the processing further comprises obtaining a media access control address from a network server of the school and storing the media access control address in a data set.
4. The method according to claim 1 , wherein the processing further comprises, responsive to the student identifier not found in the wireless local area networking data provided by the school, determining the student identifier by correlating a media access control address and timestamps found in the wireless local area networking data provided by the school with a class schedule associated with the student, the class schedule containing the student identifier.
5. The method according to claim 1 , wherein generating a forecast of future behaviors of the student comprises: constructing an autoregressive integrated moving average model; and solving the autoregressive integrated moving average model to obtain a forecast of future behaviors of the student over time.
6. The method according to claim 1 , wherein determining at-risk student retention thresholds comprises utilizing at least one of decision tree learning, logistic regression modeling, or standard deviation outlier threshold modeling.
7. The method according to claim 1 , further comprising:
generating a report containing an outcome of the at-risk student engagement detection; and
delivering the report to the client device through a student engagement application or an email.
8. A student engagement and analytics system, comprising:
a processor;
a non-transitory computer-readable medium; and
stored instructions translatable by the processor to perform:
processing wireless local area networking data into data sets having an internal data structure, the wireless local area networking data passively collected by wireless access points of a school;
analyzing the data sets to produce objective measures of student behaviors specific to a student of the school, the analyzing comprising:
correlating place and time information in the data sets specific to the student;
associating the place and time information to a student identifier of the student and to a longitude and latitude coordinate of a physical location at the school; and
generating a plurality of behavior tables, each behavior table of the plurality of behavior tables representing an objective measure of a student behavior;
performing, by the server computer, machine learning of student behaviors utilizing the plurality of behavior tables, the performing comprising:
constructing a multivariate linear regression model utilizing a set of student behavior variables;
solving the multivariate linear regression model to obtain an actual student engagement score representing an overall student retention of the student; and
generating a forecast of future behaviors of the student utilizing the actual student engagement score;
performing, by the server computer, at-risk student engagement detection utilizing the actual student engagement score, the at-risk student engagement detection comprising:
determining at-risk student retention thresholds using machine learning; and
determining an at-risk student engagement detection outcome utilizing the at-risk student retention thresholds; and
updating a user interface on a client device in real time to include the at-risk student engagement detection outcome.
9. The system of claim 8 , wherein the internal data structure is internal to the student engagement and analytics platform and has a set of attributes, the set of attributes including at least a source network address, a destination network address, and a timestamp.
10. The system of claim 8 , wherein the processing further comprises obtaining a media access control address from a network server of the school and storing the media access control address in a data set.
11. The system of claim 8 , wherein the processing further comprises, responsive to the student identifier not found in the wireless local area networking data provided by the school, determining the student identifier by correlating a media access control address and timestamps found in the wireless local area networking data provided by the school with a class schedule associated with the student, the class schedule containing the student identifier.
12. The system of claim 8 , wherein generating a forecast of future behaviors of the student comprises: constructing an autoregressive integrated moving average model; and solving the autoregressive integrated moving average model to obtain a forecast of future behaviors of the student over time.
13. The system of claim 8 , wherein determining at-risk student retention thresholds comprises utilizing at least one of decision tree learning, logistic regression modeling, or standard deviation outlier threshold modeling.
14. The system of claim 8 , where in the stored instructions are further translatable by the processor to perform:
generating a report containing an outcome of the at-risk student engagement detection; and
delivering the report to the client device through a student engagement application or an email.
15. A computer program product comprising a non-transitory computer-readable medium storing instructions translatable by a server operating on a student engagement and analytics platform to perform:
processing wireless local area networking data into data sets having an internal data structure, the wireless local area networking data passively collected by wireless access points of a school;
analyzing the data sets to produce objective measures of student behaviors specific to a student of the school, the analyzing comprising:
correlating place and time information in the data sets specific to the student;
associating the place and time information to a student identifier of the student and to a longitude and latitude coordinate of a physical location at the school; and
generating a plurality of behavior tables, each behavior table of the plurality of behavior tables representing an objective measure of a student behavior;
performing, by the server computer, machine learning of student behaviors utilizing the plurality of behavior tables, the performing comprising:
constructing a multivariate linear regression model utilizing a set of student behavior variables;
solving the multivariate linear regression model to obtain an actual student engagement score representing an overall student retention of the student; and
generating a forecast of future behaviors of the student utilizing the actual student engagement score;
performing, by the server computer, at-risk student engagement detection utilizing the actual student engagement score, the at-risk student engagement detection comprising:
determining at-risk student retention thresholds using machine learning; and
determining an at-risk student engagement detection outcome utilizing the at-risk student retention thresholds; and
updating a user interface on a client device in real time to include the at-risk student engagement detection outcome.
16. The computer program product of claim 15 , wherein the internal data structure is internal to the student engagement and analytics platform and has a set of attributes, the set of attributes including at least a source network address, a destination network address, and a timestamp.
17. The computer program product of claim 15 , wherein the processing further comprises obtaining a media access control address from a network server of the school and storing the media access control address in a data set.
18. The computer program product of claim 15 , wherein the processing further comprises, responsive to the student identifier not found in the wireless local area networking data provided by the school, determining the student identifier by correlating a media access control address and timestamps found in the wireless local area networking data provided by the school with a class schedule associated with the student, the class schedule containing the student identifier.
19. The computer program product of claim 15 , wherein generating a forecast of future behaviors of the student comprises: constructing an autoregressive integrated moving average model; and solving the autoregressive integrated moving average model to obtain a forecast of future behaviors of the student over time.
20. The computer program product of claim 15 , where in the instructions are further translatable by the server computer to perform:
generating a report containing an outcome of the at-risk student engagement detection; and
delivering the report to the client device through a student engagement application or an email.