Abstract:

Use of Educational Analytics at a large-scale represent a key advance in use of technology to enhance educational systems. Because of a variety of reasons including lack of transparency and bad governance, educational analytics can provide unique insights into improving educational systems in developing economies. Results from such analytics can be used at multiple levels within a school system. For example, teachers who are often ill-trained can be provided scaffolding to better plan lesson, or to allocate appropriate time on task for various groups of students of varying ability in their classrooms. Similarly, first-line educational administrators typically responsible for tens of schools can use this information to observe trends and patterns of weaknesses and strengths within and across their schools and to take remedial action like conducting just-in-time training sessions. At a yet higher level where administrators are responsible for overseeing thousands of schools, the same data can be fed into policy formulation. While potentially effective, such implementations of Educational Analytics must be adapted to meet the unique constraints of developing countries with respect to availability of technology and the motivation and capacity of the various stakeholders to use it. This paper presents the design and implementation of an end-to-end Educational Analytics system specifically designed for developing countries. This system is currently being deployed in a large province of a developing country in over tens of thousands of primary public schools. The system integrates Moodle Learning Management System (LMS) and the Drupal Content Management System (CMS) with a custom-designed automated assessment recording systems that generates paper-based bubble-sheets for monthly student assessments. These paper-based assessments are attempted by the students once a month for each key subject area. The results of the assessment are subsequently scanned and automatically fed into to the LMS via the Internet. The LMS, in turn, is interfaced with the Pentaho Analytics platform where assessment and teacher observation data are mapped to the various OLPC cubes; each cube represents a particular viewpoint that can be teacher-centric, school-centric or child-centric, for example. A host of reports are available online for a variety of stakeholders including teachers, principals, and educational administrators. For example, the School level reports show monthly progress of children. For each grade (e.g., Grade V) and Subject (e.g., Math) a teachers can drill down to Unit (e.g., DECIMALS AND PERCENTAGES), Topic (e.g., Multiplication and Division) and Student Learning Outcome (e.g., multiply numbers up to six digits). In addition, the teacher can view the profile of each student with respect to any of the aforementioned dimensions. Similarly, principals and first-level educational administrators can view each teachers’ ranking on a monthly basis based on a number of criteria including use of teacher diary, assigning and checking homework, activity-based teaching and learning, use of visual aids, adherence to lesson plans, quality of interaction with students, and general classroom management. At a higher level, an educational administrator responsible for hundreds or thousands of schools can view the various performance parameters of the school (e.g., teacher attendance, enrollment) or clusters of schools using a Google maps interface.

Keywords:

Educational analytics, Pentaho, developing countries.