Descriptive Design of Learning Analytics and Visualization

CHAPTER THREE

Descriptive Design of Learning Analytics and Visualization

3.1: Theories and Models Employed in the field of Learner Analytics

In any study, theories and models provide vital background information upon which users can base important decisions and arguments. Since these tend to have been tested and accredited to be true and factual, they are objective and therefore reliable. Relative information is fundamental for shaping subsequent knowledge generation in the respective field of study. Seemingly, modern researchers rely heavily on these models and theories in development of important hypotheses. In learning analytics and visualization, there are various theories, frameworks and models that inform this field of study. These include the Tanya Elias learning analytical model, collective application model and the five-step model of learning analytics amongst others. This faction of the paper reviews these models in a bid to underscore how they shape and influence the field of learner analytics.

Tanya Elias Learning Analytics Model

In her model, Tanya argues that learning analytics can be compared to the knowledge continuum that was proposed by Baker (Baker2007). At the data stage of this model, raw facts are obtained by the individual pursuing the respective analytics. This is followed by the information stage. Essentially, this seeks to define and give meaning to the data that is obtained (Ahasan & Imbeu, 2003). This is further followed by the knowledge phase. This phase is characterized by an intensive analysis and synthesis of the derived information (Snibbe, 2006).

Finally, the wisdom stage allows users to employ the respective knowledge in establishing and achieving set goals. Fundamentally, Elias believes that learning analytics constitutes of theory, people and organizations. In particular, Long and Siemens (2011) cite Elias postulating that the learning analytics employs four distinct technological resources to effectively complete a continuous pattern of three step wise cycles. The respective cycles are aimed at improving learning as well as teaching.

The Five Step Model of Learner Analytics

This model was proposed by Campbell and Oblinger. In their review, they asserted that learner analysis is a step wise process that is characterized by five systematic phases. These include ‘capture, report, act and refine’ (Campbell & Oblinger, 2009). With respect to capturing, data is derived from learning management systems or content management systems that are integrative of student information systems. Researchers at this level can also capture important additional information from historical background information of the affected learners or form vital demographic statistics.

In his research, Corbitt (2003) posits that such data is often stored categorically under classes such as residential life, financial aid, tutoring centres and so forth. In most instances, learning institutions also tend to have access to student accounts. Such data tends to be related to how the students use institutional technology or how they access important course materials. In essence, the pieces of important data are always fragmented.

Once these have been captured, they are reported relevant individuals. According to Baker (2007), these proceed to identifying hidden trends, patterns as well as possible exceptions in these data. Using this data, the relevant stakeholders can make predictions regarding students who might be at risk and develop viable intervention measures. Mostly, actions that are taken include informing relevant stakeholders and taking necessary and timely interventions accordingly.

Particular emphasis is placed on specific intervention measures to be undertaken, when to implement the respective measures and how to do and how often to develop ideal models that accurately measure student success. Finally, the characteristic continuous improvement loop defines the refinement process. This includes timely delivery of emergent data that stems from various factors such as participant behaviours, curriculum changes, process improvement and so forth (Campbell & Oblinger, 2007).

Collective Application Model

This was put forth by Dron and Anderson (Bogers & Daguene, 2003). The model seeks to define the learning analytics that are characterized by collectives, networks and groups. The model indicates that various technologies are supportive of a cyclical process. This in return supports the continued improvement model that is characterized by information gathering, processing and presentation. In the gathering stage, critical data is selected and captured. Process according to Johnson, Adams and Cummins (2012) constitutes aggregation of the respective information.

Displaying or presenting the information is defined by information sharing. Finally, dissemination includes making vital decisions based on the respective information (Goldstein & Katz, 2005). Notably, the model is systemic in nature. The respective system has intrinsic challenges as well as opportunities. The system design has varied features including the specific content. In addition, it is employed by individual users and impacts in different ways to collective users too.

The SoLAR Framewok

This model addresses advocates for an integration of four specific resources and tools. Firstly, it cites that learning analytic engine that is complex and versatile in nature. This according to Wang and Ren (2009) is useful for gathering and processing data. This it achieves through the use of diverse analysis modules. Secondly, there is the adaptive content engine. Thirdly, the intervention engine offers recommendations and provides automated support to the entire process. Finally, there is the dashboard, visualization and reporting tools.

The Open University Netherlands framework

According to this proposal, learning analytics is comprised of six critical dimensions. The first important dimension is competencies. This is defined by critical thinking and effective interpretation. There is then the constraints facet that constitutes ethics and privacy. The third dimension includes stakeholders. These are numerous and include parents, learners, teachers and the learning institutions within which the process of learning takes place. The forth components involves objectives that include prediction and reflection aspects. The fifth dimension according to Fritz (2012) includes educational data that is represented by either closed or open sources. Finally, there is technology facet that is defined by relative processes such as statistical analysis. Notably, these dimensions are structural in nature and they contribute to learner analytics in different ways.

The preceding models and frameworks offer useful insights in the understanding of learner analytics. From them, it cannot be disputed that the concept of learner analytics has distinctive building blocks that define it. Relative dimensions share intricate relationships and present the field as inherently systemic in nature. Individual dimensions contribute in different ways to the holistic whole. Summarily, an analysis of the models indicates that there are seven constituent stages of learning analytics. These include selection and capturing, aggregation, reporting, prediction, usage, refining and dissemination or sharing. In his research, Fritz (2012) indicates that web analytics on the other hand includes definition of goals, analysis of the respective goals, measuring them and sharing the outcomes. The systemic nature of the learner analytics process also underscores the fact that all constituents or facets are vitally important.

3.2: Visualization Techniques Used by Learner Analytic Tools (4pgs)

Employment of visuals in presentation of student and teacher overviews in learner analytics is of paramount importance. Visuals present a clearer understanding of various relationships that characterize the learning experience. They are appealing to the eye and aid in increasing the interest of the learner in the topic under review. Most importantly, they enhance a greater understanding of the results as well as relative implementations. One of the main characteristics of learner analytic tools includes visual techniques. Apart from learners and instructors, other stakeholders such as administrators also benefit immensely from these visualizations. Most of these have graphical representations of the intricate relationships between the learner and the educator at varied levels. This section explores visualization techniques that are used by learner analytic tools in a bid to underscore their importance in enhancing understanding and greater appreciation of outcomes.

In his review, Eckerson (2006) cites that dashboards specifically employ various forms of visualizations in presenting important data to the audience. One of the most common techniques includes charts such as bar charts. These are also referred to as histograms and underscore the distribution of the values of a single variable. The number of objects is often represented by the height of each bar. Another common type of chart is the pie chart that is represented by a circular chart that tends to be divided into different sectors. The sectors are representative of the size of every value presented in the findings. Another equally important visualization technique employed by learner analytic tools is the line graphs. Basically, this has points that are connected by lines that indicate changes that are encountered through time or when something else happens. Ideally, this is used to indicate trends in important issues such as learner progression and teacher performance. Using these, it cannot be disputed that both parties can have a better understanding of the inherent relationships.

Just like the dashboard, the social network analysis tool also employs a host of visualizations to enhance instantaneous user understanding. In his study, Norris et al (2008) indicates that it draws on varied concepts related to graph theory as well as structural theory. Moreover, this tool evaluates relative network properties including density, connectivity, centrality, degrees and betweenness. Important information is visually represented graphically to ease use by instructors as well as learners. Notably, the information presented is multifaceted and very complex. Nonetheless, graphical presentations ensure that the respective information is effectively understood by the audience. Although simple, edufeedr visualization is useful, organized and understandable (Dawson, Heathcote & Poole, 2010). It comprises of a matrix that has a row for each learner. This clearly displays the progress or performance of the learner in each examination. Using this, both the learner and the educator can be able to closely monitor trends and make sustainable interventions accordingly. Notably, the visualizations can be transferred to individual student blogs, virtual learning environments or PLE widgets that are provided by the respective institution.

In his research, Corbitt (2003) cites that use of flow diagrams enhances user understanding. This is particularly so in instances where they are descriptive. Fundamentally, flow diagrams briefly explain the relationships between different variables in light of cause effect. Using this, the learners as well as educators can be able to understand and appreciate the outcomes or consequences of their activities. Basing on the trends, negative tendencies can be avoided and positive ones reinforced accordingly. Gauges are also important visual techniques that are explored by dashboards. Dawson et al (2010) defines a gauge widget as a very simple status indicator whose display has a needle moving within different numbers that are displayed on the edges. It can be likened to a speedometer of a car. Just like thermometer as well as cylinder widgets, dashboard widgets are designed in a manner that allows them to display the value of just one metric. In this regard, the needle found within the gauge visually represents the respective single metric value. To enhance its usefulness, it is often used in combination with a selector. Basically, this gives the users a chance to choose the specific metric values that they wish displayed on the screen at a given time.

Also worth mentioning is the use of lines by tools such as the Student Activity Monitor. Relative graphical representations enhance awareness for educators and encourage self monitoring for students. Notably, most of these employ different colours that further increase their appeal. Norris, Baer, Leonard, Pugliese and Lefrere (2008) assert that insertion of recommendation areas in the respective tools enables the users to navigate to the exploration segment too. In most instances, this provides useful insights regarding what needs to be done to improve the process of teaching and learning. As visual techniques, maps are also commonly employed by dashboards. Generally, maps constitute visual representations of a given area. Symbolically, they highlight any relationships between spatial elements such as regions, themes, and objects amongst others. When used appropriately, maps illustrate learner analytics data in inviting and clear ways. The fact that they are specific and complex enhances effective representation of important data. Ultimately, they enable the users to make vital decisions regarding the outcomes of data analytics. In this regard, Trinkle (2005) believes that they are indeed important decision support tools.

According to Snibbe (2006), performance related data is best represented using score cards. This visualization technique illustrates performance metrics of both the learners and the educators. Using this, all stakeholders can be able to understand the performance of the institution or organization at a glance. Software based score cards usually place great emphasis on individual accountability. In other words, it evaluates the contributions of each individual to achieving the institutional or organizational strategic goals. Emergent research indicates that the difference between dashboards and scorecards is increasingly being blurred. This is due to the fact that the two perform almost similar analytic functions and one can be used in place of another. Nonetheless, it is worth appreciating that unlike scorecards, dashboards deal with tactical information and perform more advanced analytic roles. Besides drilling into very top level information, dashboards also analyze supporting data. Comparatively, relative software is more complex than that of a scorecard (Fritz, 2012). At this point in time, it cannot be disputed that visual techniques are an intrinsic aspect of learner analytics. Using this, one can be able to present to the audience complex information at a glance. There importance is further increased by the fact that they offer other links through which users can navigate to more detailed data. Conclusively, data presented by these techniques is not only multifaceted but also clear and inviting to the user.

3.3: Skills and Competencies for the Effective Use of Learner Analytic Tools (3pgs)

3.4: Applications of Learner Analytics in the Fields of Academics and Organizational learning

Learner analytics have various applications in both the academic and organizational spheres. Relevant stakeholders rely on it for performing useful tasks that benefit both the learners and the educators. According to Goldstein and Katz (2005), learner analytics are used for user modeling. Using this, authorities are able to determine the knowledge of the learners as well as behaviors. Furthermore, they can be able to determine the motivation degree of both the learners and the educators. Through responses to questionnaires and surveys as well as analysis of the performances of students in different units, one can be able to underscore the experience and satisfaction of the learners and educators through learner analytics.

In education, Cho, Kim and Kim (2002) posit that learner analytics is used for determining the areas that need improvement within the learning environment. This is based on the type of responses that are provided by the students about different measurable variables. Relative decisions are often based on the performance of the students in the identified fields of study. Using the current experience, one is able to identify inherent weaknesses that can be mended to enhance better performance in future. In this regard, the improvement areas benefit the next users and contribute significantly to the enhancement of quality performance. In the corporate sphere, this ensures sustained performance and ensures optimal output.

Also worth appreciating is the application of learner analytics in determining the adaptation of the users. Once the weak areas have been identified, it’s notable that not all of them can be improved upon immediately. In such instances, users need to devise alternative ways in a bid to survive and continue performing optimally. The best options in such a scenario would be to alter user experience to be in line with the current provisions. Learner analytics provide viable options through which this good can be furthered (Cho, Kim & Kim, 2002). Relative to this is the use of learner analytics to personalize the instruction methods. In his review, Baker (2007) defines academic personalization as the adaptive pacing as well as styling of the instruction approaches to be consistent with the interest and preferences of the learner. Effective personalized needs to be informed by user feedback or recommendations. These constitute the views of the students regarding how the process of instruction can be improved. For instance, students can suggest use of different content, periodic testing of their progress and more practice amongst other suggestions. Since they are user generated, they address the concerns of the users in a sustainable and effective manner.

In the academic and organizational spheres, learner analytics is used in user profiling. User profiling constitutes a collection of important personal data that describes the basic characteristics of the respective user. Essentially, this seeks to cluster the users into distinct groups and categories (Trinkle, 2005). Notably, students have varied preferences, backgrounds, interests as well as short term and long term learning goals. This is important in institutions or organizations where the student population or workforce is high respectively. This eases analysis of relative variables and makes it possible for the relevant stakeholders to draw viable conclusions in a timely and effective manner. Profiling enables the stakeholders to adapt personalized learning environments that are consistent with the needs of the profiled groups or individuals. To a great extent, this enables learners as well as educators to optimize efficiency and effectiveness.

In his research, Noris et al (2008) ascertains that profiling technologies are applicable in various domains for different purposes. In the corporate sector for instance, relative technologies generate important knowledge about the behavior of clients as well as their preferences. Using this, organizations are able to offer customized products and ultimately enhance their performance in the competitive environment. Likewise, the technologies can inform development of customized learning frameworks in the academic sphere.

An equally important application of this field of specification within the identified spheres pertains to domain modeling. Fundamentally, a domain model represents the main concepts of a topic area or subject (Mobasher, Colley & Srivastava, 2000). Domain modeling also underscores various relationships between the entire units or concepts of a particular study. In learning analytics, domain modeling reviews how the process of learning is affected by changes in the presentation sequences of the relative topics. In the academic sector, domain modeling has increasingly been used for fine tuning the process of instruction in a bid to enhance the process of learning.

Learner analytics is also used for analyzing the learning components, instructional principles and the curriculum (Wang & Ren, 2009). Particular areas of emphasis in such scenarios include determination of the most important components that are effective at enhancing the process of learning, identification of the most viable and workable principles and effectiveness of the entire curricula. Data from this is used for reviewing the quality of the education being provided to the learners. In this regard, it is worth appreciating that the preceding factors are intricate and share augmenting relationships. Most importantly, they contribute directly to improvement of capacities of students in the institutional and organizational environment.

This field of specification is also used for analyzing trends within the organizations or institutions. In particular, it determines how the change occurs, what type of change and within which specified period of time (Cho & Kim, 2004). Appropriate data in this respect is collected from completion records, enrollment records and students sources in consecutive years. The specific type of information varies and is solely based on the area of interest. Learner analytics provide the basement upon which credible recommendations regarding the most viable course of action is made. In his research, Fritz (2012) indicates that using the results, one can be able to determine the next best action of the respective user. In order to ensure objectivity, it is important to base the recommendations on the historical data or background of the user.

This section has clearly detailed the various sets of application of learner techniques in the corporate and academic sphere. The ultimate aim of this is to adopt as well as personalize a teaching and learning environment that is effective and responsive to wide ranging student and organizational needs. These applications are at the core of the objectives of learner analytics. Based on them, all stakeholders can be able to identify various ways through which they can benefit from this field of specification. Being an emergent field of study, it present a host of opportunities to the client base that await to be tapped and employed for enhancing the complex process of learning and instruction.

3.5: Benefits of Learner Analytics Tools

At this point, it is worth appreciating that learner analytic tools are useful in the academic as well as corporate circle. This is attributable to their various applications that have been reviewed in the preceding section. As it has come out, the tools can be used for intervention and reinforcement purposes and different stages of learning and teaching. Essentially, they seek to improve the process of learning and ensure that t meets the demands of the learners and expectations of the educators. From a corporate point of view, this ensures that learners are equipped with sufficient skills and knowledge that enable them to perform optimally. The fact that this is at the core of most corporate entities in this highly competitive environment cannot be disputed. This faction of the paper provides an in-depth discussion of the benefits that are provided by the learner analytics software applications. This is done in light of the importance of the respective uses in both educational institutions and corporate organizations.

Ideally, learning analytics are closely associated with sophisticated and powerful computer software (Velestsianos, 2010). Although relative analytical processes are complex, they are systematic. They are based on cognitive, social and technical processes. The analytic technology on the other hand relies on effective use of info-ware, orga-ware, human-ware and techno-ware (Astin, 1993, Bogers & Daguere, 2003). In learner analytics, techno-ware is represented by computers as well as their software. These benefit the learning institutions in different ways. To begin with, they allow learning institutions to collect important data regarding the learning as well as teaching process. After computation and analysis, this data offers useful insights into the performance of both parties. Credible decisions regarding how any weaknesses can be addressed are then made based on this data. This ensures that the ultimate decisions are based on informed thought as opposed to theoretical conceptions.

In his research, Baker (2007) asserts that the information that is usually captured by Learner Management Systems (LMSs) is complex, objective, reliable and not easily questionable. The relative data is reportedly rich in terms of quantity and diversity. In most instances, it addresses factors pertaining to learner online behavior, accessibility of students to important ICT resources and interaction of the students with the teachers as well as peers. In essence, the data gathering hardware and software do a commendable job in systematically arranging vital and relevant information during the very initial stage of analysis. The only challenges that the techno ware has experienced pertains to effective integration of the varied data sources (Snibbe, 2006).

After collecting the important relevant data, various predicting and reporting tools process the respective information. These tools are wide and varied and include decision trees, data visualization, machine learning, regression analysis, artificial intelligence and neural networks (Corbit, 2003). Visual networks are particularly useful because of their contribution to timely and informed decision making. When presented and packaged correctly, they enable the user to make critical decisions effectively. As aforementioned, these presentation techniques include maps, gauges, dials, graphs and charts amongst others. In light of education, visual representations enable the educators to interpret and evaluate the pedagogy that is in use. Emergent research ascertains that using predictive models and dash like user interfaces, both students and teachers can benefit significantly from meaningful information contained in Learner Management Systems and Content Management Systems (Campbell & Oblinger, 2007).

Nonetheless, it is worth appreciating that use of dashboards in learner analytics is compounded by various limitations. This is typical of use of any technological innovations in performing critical tasks. In particular, Goldstein and Katz (2005) argue that although the tool is visually appealing, it does not have the ability to provide the audience with information that is truly useful. Under normal circumstances, this limitation can be countered by ensuring that the source data from which the information is derived is timely, credible and accurate.

The Social Network Analysis tool is also very beneficial in the academic sphere. Using this, one can be able to not only extract and collate information but to also evaluate and visually represent relative outcomes. The fact that the process is timely ensures that any decisions or interventions are developed, implemented and enforced in a timely and effective manner. In the corporate sphere, this is instrumental in enabling an organization to maintain its competitive edge. The ability of the tool to identify the relationships between different variables enhances a more focused response. In the long run, it improves the performance of institutions and organizations and ensures that they operate in a sustainable manner.

Besides gathering and evaluating important information, Cho and Kim (2004) indicate that instructors are increasing relying on these tools during knowledge application. In this respect, Moodlog logs and Moodle logs proactively provide important feedback to their users in the learning environment. Under ideal conditions, Moodlog sends an email to the instructors and students automatically in an effort to remind them to either view or download a relevant resource. This is important in boosting the performance of both parties in the learning environment. Arguably, the practice is informative and contributes a great deal to education and capacity building.

These tools also encourage accountability especially in service delivery. According to Corbit (2003), this is attributable to their ability to track the performance of students and teachers as well as employees in organizations. Teachers use the data regarding trends to improve on their instructional methods and customize the same to meet student needs. The students on the other hand use the performance trends to identify their weak areas and develop ways through which they can address these in a timely and efficient way. In the corporate environment, review of performances and promotion of accountability benefits the organization in different ways.

Just like academic institutions, corporate entities can use data related to performance trends to identify it’s failing areas and develop credible intervention measures. By being able to track employee performance, organizations can quantify the contributions of individual employees to company growth and development, identify and address the specific needs of the employees and finally enhance its overall performance in the market environment. In sum, learner analytic tools contribute positively to the growth and development of organizations and to the performance of learning institutions. As it has come out from the preceding review, the tools perform various functions and generate important information. This information is used for critical decision making as well as knowledge application. Use of these tools encourages accountability that sets the pace for capacity building and improved organizational and institutional performance. Put differently, the tools enable the organizations and institutions to meet their goal and objectives.

References

Ahasan, R. & Imbeu, D. (2003). Socio-technical and ergonomic aspects of industrial technologies. Work Study, 52 (2/3), 68-75

Astin, A. (1993).What matters in college? Four critical years revisited. San Francisco: Jossey-Bass.

Baker, B. (2007). A conceptual framework for making knowledge actionable through capital formation. USA: Maryland University College

Bogers, M. & Daguere, M. (2002). Technology transfers in international joint ventures. University of California: Berkeley.

Campbell, J. & Oblinger, T. (2009). On the design of collective applications. In Proceedings of the 2009 International Science and Engineering, 4, 368-374.

Cho, Y., Kim, J. & Kim, S. (2002). A personalized recommender system based on web usage mining and decision tree induction. Expert Systems with Application, 23 (3), 329-342.

Cho, Y. & Kim, J. (2004). Application of web usage mining and product taxonomy to collaborative recommendations in e commerce. Expert System with Application, 26, 233-246.

Corbitt, T. (2003). Business intelligence and data mining. Management Services, p. 18.

Dawson, S. Heathcote, L. & Poole, G. (2010). Harnessing ICT potential: The adoption and analysis of ICT systems for enhancing the student learning experience. International Journal of Educational Management, 24 (2), 116-128.

Eckerson, W. (2006). Performance dashboards: Measuring, monitoring and managing your business. Hoboken, NJ: John Wiley & Sons.

Fritiz, J. (2012). Classroom walls that talk: using learning analytics to raise awareness of underperforming students. Presentation at Symposium on Learning Analytics at Michigan

Goldstein, P . & Katz, R. (2005). Academic analytics: The uses of management information and technology in higher education. Educause Centre for Applied Research, 8.

Johnson, L., Adams, S. & Cummins, M.