ANLY482 AY2016-17 T2 Group15 Project Overview
In today’s educational world, more and more educational institutions are incorporating technology into teaching and learning to enrich students’ learning experiences and improve teachers’ pedagogical practices. The dilemma that many secondary schools face is how to aptly establish the right criteria to recommend the right subject combinations to students, so as to improve their learning outcomes. For instance, it is difficult for teachers to decide whether or not to recommend students to take on the subject combinations of Double Science or Triple Science. Should schools determine the capability of students based on their overall examination grades, or should they base their decisions on their individual subject grades (such as Mathematics or Science)? Often, many parents believe that their child is capable of coping with Double or Triple Science combinations, even if their Secondary 2 results show otherwise. Without proper analytical evidence, it is difficult for teachers to recommend students the right subject combinations that would ultimately improve their 'O' Level performance.
Edufy Secondary School is a heartland neighbourhood secondary school located in the North region of Singapore. The school is committed to providing an ideal learning environment and experiences for its students. Despite having a comprehensive set of past students’ data, the school lacks the expertise to analyse the data in a way that can aid in their decision making.
The role of data analytics is becoming even more relevant and important, given the rise of Learning Analytics. Learning analytics seek to improve teaching and learning through the targeted analysis of students’ academic performance data [1][2]. By analysing the past data of students’ examination results using various data analysis and visualization techniques, it enables the school to discover useful patterns and relationships within the data. These insights equip the school with the intelligence that would enable them to better understand students’ performance and make informed decisions in their curriculum to refine their pedagogical strategies and optimize student performance [3].
Utilizing past data of students’ grades from the school’s database, we aim to discover useful and practical insights which will allow teachers to better decide and advise students on choosing their Secondary 2 subject combination, particularly on whether they should take one of the following subject combinations:
- Combined Science
- 1 Pure Science and 1 Combined Science
- Double Pure Sciences or
- Triple Pure Sciences
We will attempt to analyze the trends of students' academic performance by examining their past subject grades and subject combinations.
To achieve the above mentioned, we have performed an in-depth analysis on the historical data with the following aims:
- To help secondary schools and teachers better formulate the right streaming practices and criteria that would benefit all students
- To develop an application using R for the school so that they can input future data to improve the accuracy of the model in predicting students’ GCE ‘O’ Level examinations results
In recent years, there has been the development of several dashboard applications to support teaching and learning in the education sector. Particularly, the advent of learning dashboards help teachers improve their knowledge of students by providing tools for the review of analysis of students’ history [4]. Such dashboards often provide teachers with the graphical representation of a student or a course, often in the form of bar charts and matrices. Such visualization techniques enable them to make flexible, data-driven decision making. While previous research focuses on the evaluation of course activity and teaching practices, none has been done specifically in analysing students’ examination scores. As such, our study will attempt to fill in the gap and shed light on how learning dashboards can be applied to generate insights of students’ academic performance, thereby improving their learning outcome.
However, the challenge lies in visualizing and generating insights from the large datasets and sources. To tackle this problem, we explored the use of two visualization methods: box-and-whisker plot and tableplot. The boxplot is a simple yet powerful tool for displaying a single group of data, allowing the user to easily study the summary of the distribution [5]. On the other hand, the tableplot can display the aggregated distribution patterns of numerous variables in one single figure [6]. The tableplot is a valuable tool for inspecting statistical data, especially when there are numerous variables and a large dataset involved. We will demonstrate how the use of these methods can aid teachers in analysing students’ performance in our paper and through our application.
The figure above shows the data methodology process that we have adopted for our study. The dataset comprises of 3 batches of historical students’ data from our project sponsor, consisting of students who took their GCE ‘O’ Level examinations and graduated in years 2014, 2015 and 2016. The student records comprised of their respective subject results for each of the continual and semester examinations (i.e. CA1, SA1, CA2 and SA2) from Secondary 1 to 4, as well as their PSLE and GCE ‘O’ Level results. To protect the confidentiality of students and to ensure accuracy of the results, the names of students have been coded by our Project Sponsor.
[1] Elias, T. (2011). Learning analytics: Definitions, processes and potential. Unpublished Internal Whitepaper of Athabasca University, Canada. Retrieved from https://landing.athabascau.ca/ mod/file/download.php?file_guid=43713
[2] Fritz, J. (2010). Classroom walls that talk: Using online course activity data of successful students to raise self-awareness of underperforming peers. The Internet and Higher Education, 49, 89–97.
[3] Larusson, J., White, B., & SpringerLink. (2014). Learning analytics : From research to practice , New York : Springer.
[4] Haythornthwaite, Caroline, De Laat, Maarten, Dawson, Shane, Verbert, Katrien, Duval, Erik, Klerkx, Joris, . . . Santos, José Luis. (2013). Learning Analytics Dashboard Applications. American Behavioral Scientist, 57(10), 1500-1509.
[5] Benjamini, Y. (1988). Opening the Box of a Boxplot. The American Statistician, 42(4), 257-262.
[6] Martijn Tennekes, Edwin De Jonge, & Piet J. H. Daas. (2013). Visualizing and Inspecting Large Datasets with Tableplots. Journal of Data Science, 11(1), 43-58.
Besides the team and our supervisor, the other stakeholders are:
Sponsor
- Mr Lee Peck Ping, Principal of Edufy Secondary School (ESS)
- Mdm Lim, Vice Principal of ESS
Other Stakeholders
- Students of ESS
- Teachers and Heads of Department (HODs) of ESS
- Parents of students studying in ESS
