Keywords

1 Introduction

Smart Education (SmE), Smart University (SmU), Smart Learning (SmL), and Smart Classroom (SmC) concepts are rapidly gaining popularity among the world’s best universities because modern and sophisticated smart technologies, smart systems, smart devices, as well-data-driven and data-based strategies and solutions, create unique and unprecedented opportunities for academic and training organizations. Using those innovative concepts, universities and colleges may obtain higher standards and innovative approaches to (1) education, learning, and teaching strategies, (2) modeling of students/learners as objects and education/learning as processes, (3) unique and/or high technology-based services to local in-class and remote/online students, (4) set-ups of modern highly technological smart classrooms with easy Web-based audio/video interactions between local/remote students and faculty, and collaboration between in-class and remote students, (5) design and development of Web-based rich multimedia learning content with interactive presentations, video lectures, Web-based interactive quizzes and tests, instant knowledge assessment and automatic posting of attendance, class activities, and learning assessment outcomes on course web sites, visualization of data in various forms including student, faculty and department dashboards, and many other advantageous features [1, 2].

Our vision is based on the idea that SmU, SmE, SmC, SmL – as smart systems – should implement and demonstrate significant maturity at various “smartness” levels or smart features, including (1) adaptation, (2) sensing (awareness), (3) inferring (logical reasoning), (4) self-learning, (5) anticipation, and (6) self-organization and re-structuring [3]. This is the reason that we consider emerging learning analytics (LA) as an integral part of SmE, SmL, and SmC concepts.

The Society for Learning Analytics Research defines LA as: “… the measurement, collection, analysis and reporting of data about learners and their contexts, for purposes of understanding and optimizing learning and the environments in which it occurs” [4].

“A recent report from the U.S. Department of Education makes the point that on the program and institutional level, learning analytics can play a role that is similar to that of already existing business intelligence departments and applications. Just as business intelligence may utilize demographic, behavioral and other information associated with a particular enterprise and its customers to inform decisions about marketing, service and strategy, learning analytics promises to do something similar in educational terms” [5].

1.1 Learning Analytics’ Goals and Objectives: Literature Review

Various authors of available publications define the goals of LA in different ways. For example, in accordance with Siemens [6], “The broad goal of learning analytics is to apply the outcomes of analyzing data gathered by monitoring and measuring the learning process, as feedback to assist directing that same learning process. …Six objectives are distinguished: predicting learner performance and modeling learners, suggesting relevant learning resources, increasing reflection and awareness, enhancing social learning environments, detecting undesirable learner behaviors, and detecting affects of learners.”

Suchithra et al. [7] believe that “The main purpose of Learning Analytics is to improve the performance of learners. Also, the environment of learning in which the learner undergoes is enhanced which will ultimately result in a quality education. Learning Analytics helps educator/teacher to understand the students. Learning capabilities can be improved for the learners. … The Learning Analytics aims at the curriculum design, predicting the students’ performance, improving the teaching learning environment, decision support system for Higher Education Institutions, personalized approach to individual students, online and other learning modes including mobile, subject wise teaching and learning, subjects which has practical and evaluation process in the education system. … Learning Analytics is about the collection, analysis of data about the learners. It is an emerging field in research which uses data analysis on every tier of educational system”.

Borkar and Rajeswari presented the following approach in [8]: “Learning analytics approaches in general offer different kinds of computational support for tracking learner behavior, managing educational data, visualizing patterns, and providing rapid feedback to both educators and learners”.

Additionally, Tempelaar et al. [9] argue that “The prime data source for most learning analytic applications is data generated by learner activities, such as learner participation in continuous, formative assessments. That information is frequently supplemented by background data retrieved from learning management systems and other concern systems, as for example accounts of prior education”.

1.2 Learning Analytics on University Level: Literature Review

In accordance with Suchithra et al. [10], “Institution can use academic analysis to know the success of the students. It can also be used to get the attention of public. Report of the analysis can be used for the publicity of the institution.”

Mattingly et al. [11] argue that “Learning analytics in higher education is used to predict student success by examining how and what students learn and how success is supported by academic programs and institutions. … The focus is to explore the measurement, collection, analysis, and reporting of data as predictors of student success and drivers of departmental process and program curriculum”.

Siemens et al. in [12] wrote: “It is envisaged that education systems that do make the transition towards data-informed planning, decision making, and teaching and learning will hold significant competitive and quality advantages over those that do not”.

1.3 Learning Analytics on Course Level: Literature Review

Multiple publications are available regarding the approaches, concepts, and proposed framework for LA at the course level. For example, Dietz-Uhler and Hurn in [13] argue that “Goals that learning analytics address include predicting learner performance, suggesting to learners relevant learning resources, increased reflection and awareness on the part of the learner, detection of undesirable learning behaviors, and detecting affective states of the learner. … Data as login frequency, site engagement, student pace in the course, and assignment grades to predict course outcome. … Performance on course assignments and tests at various times in the course significantly predicted final grades.”

In accordance with Dyckhoff et al. in [14], “Masses of data can be collected from different kinds of student actions, such as solving assignments, taking exams, online social interaction, participating in discussion forums, and extracurricular activities. This data can be used for Learning Analytics to extract valuable information, which might be helpful for teachers to reflect on their instructional design and management of their courses.”

Ruiperez-Valiente et al. in [15] present the following vision of learning analytics: “The Khan Academy…platform provides an advanced learning analytics module with useful visualizations for teachers and students…the Khan Academy platform provides different learning analytic features by default.”

1.4 Levels of Learning Analytics: Literature Review

In general, LA may have several levels or layers of hierarchy and/or maturity. For example, Siemens et al. in [16] introduced a general hierarchical framework for LA levels that include 3 levels for LAs: (1) LA on personal level (analytics on personal performance in relation to learning goals, learning resources, and study habits of other classmates), (2) LA on course level (social networks, conceptual development, discourse analysis, “intelligent curriculum”), and (3) LA on departmental level (predictive modeling, patterns of success/failure). Additionally, for academic analytics they introduced (4) LA on institutional level (learner profiles, performance of academics, knowledge flow, resource allocation), (5) LA on regional level (state/provincial): comparisons between systems, quality and standards, and (6) LA on national/international level.

On the other hand, Lynch et al. in [12] proposed an LA sophistication model that contains the following LA maturity levels: (1) awareness, (2) experimentation, (3) organization, students, faculty, (4) organizational transformation, and (5) sector transformation.

1.5 Smart Learning Analytics: Literature Review

The idea of smart learning analytics (SLA) is in an embryonic state at this moment; a thorough search on the Internet discovers a few relevant publications. For example, Giannakos et al. in [17] defined “Smart Learning Analytics as a subset of learning analytics that focuses on supporting the features and the processes of smart learning”. The authors primarily concentrated on “recent foundations and developments [of Smart Learning Analytics] in the area of Video-Based Learning” [16]. On the other hand, Boulanger et al. introduced in [18] “… a framework called SCALE that tracks finer level learning experiences and translates them into opportunities for custom feedback. … Students have been provided with customized feedback to optimize their learning path in programming”.

Multiple publications are available on various topics related to LA aspects. Unfortunately, the aforementioned and multiple additional analyzed publications do not provide detailed information about SLA from the smartness levels point of view, i.e. levels of (1) adaptivity, (2) sensing, (3) inferring, (4) anticipation, (5) self-learning, and (6) self-organization [1,2,3]. Additionally, the analyzed publications are focused on applications of LA to learning process of students and/or life-long learners; however, they do not emphasize the fact that LA should demonstrate “smartness” features (or be smart) and strongly support all designed smartness levels, including the “self-learning” level, i.e. an ability of a smart university “to learn” about itself and, therefore, be able “to self-optimize” its operation and main business functions.

2 Project Goal and Objectives

The overall goal of the on-going research, design, and development project at the InterLabs Research Institute at Bradley University (Peoria, IL, USA) is to use a systematic approach to identify, analyze, test, design, and eventually implement various components of SLA system for an entire SmU. In order to achieve this goal, the project team selected the following objectives:

  • analysis of most recent innovative developments in LA and SLA areas;

  • analysis of existing LA levels;

  • analysis of available software systems that may support learning analytics, and potentially, SLA;

  • identification of main features of SLA – types of data to be collected and processed, and main functionality of SLA system

A summary of up-to-date project findings and outcomes is presented below.

3 Smart Learning Analytics: Hierarchical Levels

Our vision of an SLA system is based on the concept that SLA should have a hierarchical layered structure and strongly support all major components of SmU, including

  1. (1)

    SmU stakeholders, including students, faculty, professional staff, administrators, life-long learners, donors, alumni, etc.;

  2. (2)

    SmU main smartness features, including adaptation, sensing, inferring, self-learning, anticipation, self-optimization or re-structuring (see Table 1 below);

    Table 1. SmU smart features [1]
    Full size table
  3. (3)

    SmU curricula, i.e. a set of smart programs of study and smart courses at SmU – those that can, for example, change (or optimize) its structure or mode of learning content delivery in accordance with given or identified requirements (due to various types of students or learners);

  4. (4)

    SmE and SmL at SmU – main processes and business functions at SmU;

  5. (5)

    Smart Pedagogy (SmP), i.e. a set of modern pedagogical styles (strategies) to be used at SmU;

  6. (6)

    smart learning environment at SmU, including smart classrooms, smart labs, smart departments and smart offices, etc.;

  7. (7)

    smart software systems at SmU, i.e. a set of university-wide distinctive smart software systems at SmU – those that go well beyond those used at a traditional university;

  8. (8)

    smart hardware at SmU, i.e. a set of university-wide smart hardware systems, devices, equipment and smart technologies used at SmU – those that go well beyond those used at a traditional university);

  9. (9)

    smart technology, i.e. a set of university-wide smart technologies to facilitate main functions and features of SmU and smart campus, for example, Internet-of-Things, cloud computing, iSafety, ambient intelligence, etc.;

  10. (10)

    SmU resources, including financial, technological, human, and other types of resources.

As a result, we proposed the following hierarchical levels of SLA system:

  1. (1)

    Personal level is the lowest level of the SmU for various types of stakeholders – students, faculty, professional staff, administrators, long-life learners, etc.; for example, students can get data-driven dashboard on (a) current academic performance in a course, on programs of study, etc., and (b) progress of development of various skills, including analytical, technical, management, and communication skills;

  2. (2)

    Course level for students, learners, faculty, department chair, etc.; for example, faculty can compare academic performance of a selected student with (a) all students in the same class during current semester, (b) with all students in this class during recent semesters – based on these data a faculty can predict student’s learning ability, final grade, etc.; additionally, faculty can compare student academic performance in courses that use (a) various modes of learning content delivery (face-to-face, online, blended), (b) various learning strategies (learning-by-doing, games-based learning, flipped classroom, adaptive learning, etc.)

  3. (3)

    Concentration/minor program level (i.e. a level of a group of specialized courses) for students, faculty, administrators; for example, a department chair can assess academic performance of (a) majors from his/her department, and (b) students from other departments who take a selected concentration, certificate or minor program at his/her department;

  4. (4)

    Departmental/program of study/curriculum level for students, faculty and administrators; for example, chair of department can perform predictive analysis for various students and faculty; identify patterns of success for various types of students and faculty; perform data-driven control of enrollment into departmental programs - certificate, concentration, minor and major programs;

  5. (5)

    University level is the highest level of a SmU for administrators, alumni, donors, etc.; for example, provost’s office can monitor (a) student academic performance in various courses and programs, (b) retention rate in a major, department, college, university, etc.

4 Smart Learning Analytics: Types of Data to Be Collected/Processed

An SLA system should collect numerous pieces of data in order to support the idea of the Data ↦ Information ↦ Knowledge ↦ Smartness continuum at a SmU. Several examples of types of data to be collected by SLA system are presented in Table 2.

Table 2. Types of data to be collected by SLA system at SmU (examples)
Full size table

5 Smart Learning Analytics: User Requirements

The introduced concept of SLA hierarchical levels and types of to-be-collected data enabled us to create a comprehensive list of user requirements to SLA system functionality. Table 3 below contains examples of such requirements from students, faculty, and department chair for one hierarchical level – LA on course level.

Table 3. SLA course level: SmU user requirements (examples)
Full size table

6 Analytics Systems Analyzed

Multiple software systems used in the LA area were reviewed or analyzed with the purpose of identifying those suitable for SLA; the obtained outcomes are presented in Table 4.

Table 4. Existing analytics systems reviewed or analyzed (non-comprehensive list)
Full size table

7 SLA System’s Prototype Developed

We developed a prototype of the SLA system for a smart university – the InterLabs SLA system. This system should eventually include the functionality of SLA for SmU as described in Tables 2 and 3. Graphic user interfaces of developed prototype of the InterLabs SLA system for various users are presented in Fig. 1 (student view), Fig. 2 (faculty/instructor view), and Fig. 3 (department chair or administrator view).

Fig. 1.
figure 1

A prototype of the InterLabs smart learning analytics system: student view

Full size image
Fig. 2.
figure 2

A prototype of the InterLabs smart learning analytics system: faculty view

Full size image
Fig. 3.
figure 3

A prototype of the InterLabs smart learning analytics system: administrator view

Full size image

8 Conclusions. Next Steps

Conclusions.

The performed research and analysis, as well as the obtained findings and outcomes, enabled us to make the following conclusions:

  1. (1)

    Leading academic institutions all over the world are investigating ways to transform a traditional university into a smart university and benefit from the advantages of smart university, smart classrooms, and smart pedagogy. SLA systems will play a crucial role in SmU success.

  2. (2)

    It is necessary to identify the main components of SLA systems, including hierarchical levels of SLA system, architectural model of SLA system with main SLA components and relations between them, types of data (data objects) to be collected and processed at SLA systems (Table 2), a set of SmU user requirements to SLA systems on each hierarchical level (Table 3), inputs to and outputs from SLA system, interfaces and protocols to be used, and constraints/limits of SLA system at SmU.

  3. (3)

    We analyzed 30 + current systems relevant to various aspects of SLA; unfortunately, we could not identify any mature system among the existing ones – a system that will address the SLA features and functionality as required by SmU (Table 3). As a result, we designed and developed a prototype of SLA system - InterLabs SLA.

Next Steps.

Based on the obtained research findings and outcomes, the next steps of this research project are to (a) use the Agile Software Engineering process model, (b) create a set of prototypes of SLA system for various types of users, and (c) involve various SmU users into the SLA design and development to ensure quality of final software system.