Self-Regulated Learning Platform Based on Learning and Health Data

(1)

Self-Regulated Learning Platform Based on Learning and Health Data

Md Jahedur Rahman

Master's

thesis

School of Computing

(2)

UNIVERSITY OF EASTERN FINLAND, Faculty of Science and Forestry, Joensuu School of Computing

Computer Science

Opiskelija, Md Jahedur Rahman: Self-Regulated Learning Platform Based on Learn- ing and Health Data

Master’s Thesis, 66 p., 1 appendix (3 p.)

Supervisors of the Master’s Thesis: Solomon Sunday Oyelere, Postdoctoral research- er

December 2020

Abstract: Self-direction skills are an embryonic need to sustain a healthy lifestyle along with analytical ability. Though technology promotes self-regulated learning and health monitoring, Meta-skills development not adequately studies for self- direction processes. This research aims to provide technological support for self- directed learning practices with the aid of Learning Analytics (LA) and Quantified Self (QS) based research. In this context, technological support ensures by develop- ing a dashboard with learning and health data according to the GOAL system.

In order to test the hypothesis that technological support for self-directed skills im- proves learner success, an online survey and an interview perform across Finland at different educational institutions. There were multiple choices and short questions to explain the views of the respondent. The result supported the hypothesis: technical support for self-directed skills makes learners self-management, self-confident, and helps develop skills more effectively.

Results support that health data inclusion with academic work helps maintain learners' entire learning environments to be efficient in education and life. Self-direction skill developments impact education; for a stress-free self-managing education, self- direction skill developments should be considered vital parts of education. We should need to increase more research to provide technical support in this field.

Keywords: Self-directed skill, Meta-skill, technology, Learning Analytics, Goal CR Categories (ACM Computing Classification System, 1998 version): Computer Uses in Education, Computer, and Information Science Education.

(3)

Acknowledgment

This thesis has done at the School of Computing, University of Eastern Finland, during the spring of 2020.

I want to thank my thesis supervisor, Dr. Solomon Sunday Oyelere, without whom this study was impossible. Thank you for all of your assistance and support over the years.

I would also like to thank the SELI team for their support, help, and cooperation.

It is a pleasure to thank the administration of my university for their support. I would particularly like my course coordinator, Dr. Oili Kohonen. I want to thank you for all kinds of supports.

Besides, I would like to thank my parents for their excellent advice and their kind heart. You're for me always. Finally, I could not manage my dissertation without the support of my siblings, friends, and teachers.

(4)

List of abbreviations

SDS Association for Computing Machinery ACM Association for Computing Machinery LA Learning Analytics

UEF University of Eastern Finland QS Quantified Self

OAuth Open Authorization

CSE Computer Science and Education

SELI Smart Ecosystem for Learning and Inclusions

(5)

1 Introduction

A learner's everyday lifestyle is bound to academic study, while systematic studies recognize several different practices. Healthy lifestyle is an embryonic need to control personal skills and set a multidimensional plan for managing those skills. A self- directed learner can achieve material association, processing powers, and self- judgmental abilities. Very few students can reach this ability without the help of a third party (Person) or sometimes intelligent machines. In computer science, the computer might offer several self-direction techniques (Lieberman & Linn 2014).

Technological developments expand the need for self-directed skills to plan a controlled way to develop personal goals with necessary reinforcements (Zimmerman, 2015).

Whither, Self-directed learning is a metacognitive process to develop individual learning skills by set up a goal, making a learning plan, then re-plan if the reflection is not satisfactory. The method increases one personal learning ability, predication ability, controlling their actions, and justifies works. This concept considers the most valuable framework of the 21st century, even in the healthcare fields. "Prevention is the cure," A controlled lifestyle can prevent a lot of chronic diseases. Observation behaviors collect information, and an overview for conclusion according to a particular goal is an inspiration for self-regulation (Clark, 2003).

There needs to consider a meta-skill development concept for self-regulation, which empowers other skills. But due to minimal research on self-regulation, meta-skill developments, joint action on self-direction limiting its scopes. This work aims to develop meta-skill (Hinshaw, 1991) to check personal action, reflection, and outcomes, both learning and health. According to the DAPER (data collection-analyze plan-execution monitoring-reflect) model, this thesis presents a Goal-dashboard (Majumdar et al., 2019). Those who are using this platform can create a plan for both

(8)

study duration, progress, online evaluation, comments, and feedback-based learning.

The action plan needs to improve personal activity; this model targets SELI project learning data to design a platform to show performance data and goal data. Analysis of the outcome and offers the reflection, thus watch personal gains and leakage.

We know good fitness consists of daily exercise, adequate sleep, blood pressure regulation, pulse, and calorie intake by benefit. With this data, google makes a google fitness API system where all data can find, and a personal goal system helps to improve them in the desired amount. We consider Google fitness API for health and wellbeing data.

These works make a bridge between self-aware active learning and goal bases improvements. The concept Goal is to develop from the self-regulation prediction model, where users develop their planning, execution, feedback, and re-plan process.

1.1 Research Aim

The primary goal of this work was to improve Self-direction skills (SDS) with an automated dashboard. Develop skills of self-control to regulate academic and nonac- ademic skills, according to a goal model. Our target is to collect users learning and health data, visualize them inside the dashboard and analyze to generate a plan to improve those. Besides, the learner can monitor personal actions and decide if any re-plan necessary to achieve desired reflections. This project should help a learner be independent by enhancing one's prediction powers, self-control skills, and self- analyzing skills.

In this project, a learner has a visual outcome of daily activities and reflation according to their own plan sets. A learner can find their academic and health data individ- ually and plan for both separately. Our planned dashboard makes users updated with the task performed or should need to perform. According to the desired goal, users can even re-plan their targets.

(9)

1.2 Problem Definition

Self-direction skills (SDS) are the most necessary skill for the students to support a pressure-free healthy, organized life. Students perform multiple tasks besides their academic improvement studies; those data are equally important (Majumdar et al., 2018). But It is a matter of sorrow that there is minimal research on SDS. A right study environment required physical terms like good sleep, proper physical activity, and enough food besides study. As a full package of a learner's regular activity, both skills need adequate planning for improvements.

However, this is harder to improve those skills without analysis and monitor. Col- lected data needs a proper representation of a learner so that they can identify limitations and can plan. Goal achievements need a good plan, but a plan not always be perfect. So, proper monitoring is also necessary for determining the plan was right or their needs any correction/re-plan. Finally, reflection visualization is a real-time improvement to understand outcomes.

Besides, regulation in action means analysis, plan, monitoring activities. Activities can be academic, else, or both educational and else. Here need a platform where we can plan for personal improvements. This work designs a dashboard that able to show study vs. goal and health vs. goal data. Users can plan for study and health, and then this data appear in the dashboard as goal data and compare with their academic and health data. They analyze once daily performance to illustrate reflections. This project combines data from google fit and SELI platforms. For the student who is going for SELI, courses can improve their way of learning.

1.3 Research Objectives

The main objective of this research is below:

(10)

Objective 03: Study meta-skill development process, effects over the Goal system implement process.

1.4 Research Questions

This thesis conduct based on the following questions:

RQ 01: What are the purpose of endorsing technologies to support self-directed be- haviors before exposure to the GOAL system?

RQ 02: What are the purposes to indulge in a health-enhancing lifestyle?

RQ 03: How can meta-skill development help to achieve a personal goal with aca- demic and health data?

1.5 Methodology

After analyzing 74 student data from the SELI database, see how their data can compare with their goal data. A qualitative resect approach was chosen for data arrangements and analysis. But the patterns of this research show a design-based method (Jivet, 2016). In this work, personal reflection has been evaluated based on the user plan and performance. Although the Goal dashboard focused on Self-direction skill improvements, this work also showed goal generation's design process and notified reflection. We had data from the Smart Ecosystem for learning and Inclusion (SELI) and Google fit. The learning data from SELI taken from students who took courses in that platform. Health data collected from google wearable devices or mobile phones.

For getting data, a RESTful API was designed, which could perform data GET and POST operations. We collected learning data- Activity performs (Both task complete and incomplete), Deadline, the course duration, and the courses' progress. In the case of health, we have sleeping, walking, and calorie burning data. Inside a cloud platform (MongoDB), data was stored for further operations. But, the principal activity was between data generated by the user and data carried out by them. Then, well- organized feedback provided a comprehensive comparison of user plans and user

(11)

learners' life. Finally, the outcome was visualized with charts to compare results and effects on attendees.

1.6 Chapter Summary

This chapter briefly discussed the central theme of the thesis target, problems, aims, achievement, and benefits of self-direction skills developments—furthermore, data collection, and representation process. In section 1.1 represents the research aim and outcomes of this work. Section 1.2 is an exhibit problem definition and necessity of this project. The main goals of this work clarify three fundamental aims (Section 1.3). This research occurs based on multiple questions (Section 1.4) that conduct the thesis. Section 1.5 method for this work shows how research conducts, the model's structure, model design, and the data arrangements process.

(12)

2 Literature Review

This research performs based on concepts and research about Self-direction skills, Goal concept, and self-directed dashboard. At the beginning stage of the research works, try to evaluate different research themes, patterns, contradictions, and lacking.

In this chapter, we try to extract knowledge about self-direction and its effects on learning and health. There also discussed the earlier learning dashboard the findings and lacking. Finally, we extract the benefits of self-regulation.

2.1 Meta-cognitive for Self-direction

Self-direction is a process of personal skill control in a systematic way. It is a goal- based method for every person to explain improvement through reflection (Mohd et al., 2019), where a person, by himself or with the help of something, sees their needs, builds a goal, and in the end, evaluates their self-achievement (Cazan & Schiopca 2014).

Besides, metacognition is a process of "thinking about thinking" (LDschool, 2020).

But elaborately, it is a process of judgment not only during the performing task but also plans before and after completing the task. It is a process of controlling one action consciously and conducting active research over it for planning further required adjustments on it (Keestra, 2017). A metacognition study illustrated it is a meaning- ful approach to problem-solving but does not for problem understanding (Gurat &

Medula 2016).

Metacognition strategy and self-direction are correlated terms for the students inside or outside organization revealed for effective and independent education (Cera et al., 2013). A complete relation between Metacognitive strategies with self-directed shown in (Ghomi et al., 2016) article; this was a gender-based approach to show significant differences in adopting meta-cognition. Results show no problem with adopting meta-cognitive strategies and self-management between males and females, but self-control has a huge gap with desire to learn. After the study, they come to a

(13)

decision desire to learn and self-control directly and most strongly attached with meta-cognitive strategies.

Meta-cognitive strategies help the student to develop a proper and most suitable plan of learning. These can help learners a list of activities, which can directly lead a learner toward the process of "how they have to learn." Thus, a learner becomes an independent, self-efficient, self-managing and self-directed learner. The cognitive way helps them to make a personal to-do list and go ahead with their tasks. As a result, a learner might gain full control over their activity and process (ISN, 2015).

2.2 Existing Learning Dashboard

In the 21^st century, the dashboard plays a significant role in visualizing, analyze similar or different data. In terms of data or database analysis use of dashboards is in- creasing exponentially. The most applicable definition of dashboard given by (Few, 2013), "a visual display of the most valuable information needed to achieve one or more objectives that have been consolidated on a single computer screen so it can be a monitor at a glance."

There are a few mostly used learning dashboards discussed. StepUp! is an activity tracking (Santos et al., 2012) dashboard. Actively working for self-reflection demon- strations and perform necessary comparison operations between the learner. This dashboard populates learner time spent on different tools during course assessments and tracks active presents on the forums. It shows a dynamic effect on student performance by comparing themselves with others. However, a student does not have the option to set a goal and work on self-regulation.

Classroom salon is a social learning platform. In this platform, teachers and students make a positive collaboration by (Gunawardena & Barr, 2012) creating a social network. Inside the venue, teachers teach and assess courses. Salon learners can perform

(14)

Course signals show performance with the distinct color signals, which used red, green, and orange color as signals (Arnold & Pistilli, 2012) for presenting the performance. Course track learner records like gain points, attendance, and performance. If performance below the satisfactory level, it shows a red signal. Whereas the orange signal means the position is nearby an adequate level, but improvement is needed. Whereas green says performance is satisfactory, and all are going well.

LAPA (Learning Analytics for Prediction and Action) is a virtual learning (Park &

Jo, 2015) platform. This project count student session active on the platform, session staring, and frequencies. LAPA conducts time management and self-regulation skills base performance. But the author's study result showed its focus was on personal activity management ability while learning had minimal priority.

Moodle dashboard is one of the most equipped dashboards of this decade. It runs the course, assessments (Podgorelec & Kuhar, 2011), and adds students to the forum to get feedback. Its helps to compare personal score in a pair.

Student Activity Meter (SAM) is like Moodle. Which shows activity perform during the course (Govaerts et al., 2012) and score then minimum, average, and maximum range. Like Course signals (Arnold & Pistilli, 2012) dashboard, its signal is color base, but they use gray.

There are so many more dashboards that may target the cognitive approach, such as Papanikolaou's dashboard (Papanikolaou, 2015). Some other-focused self-awareness,

‘Tell Me More,’ is one of them (Lafford, 2004). However, self-direction can bring a fundamental change in learning (Majumdar et al., 2018), which was missing or limited in the above research.

2.3 Goal System and DAPER Model

This DAPER model is developed by (Majumdar et al., 2018). In this model, DAPER means data collection, data analysis, planning, monitoring, and reflection. They plan to arrange data in the first stage (collection stage). Collected data consists of both physical and learning data. The analyzing stage has some meaning full works done to

(15)

sonal goal to execute according to the plan. The monitoring stage can help the user to show progress and re-plan according to needs. And Finally, seeing all executed processes and outcome judgments in the reflection stage.

Figure 1: DAPER Model.

The goal system used (Flanagan & Ogata, 2017) in their LMS system generates UUIDs for the user for data protection and analysis of the data privately. This project was about learning management, but its full focus on data privacy. According to this work and based on the DAPER model (Majumdar et al., 2018) makes a Self-Directed Skills base dashboard model. Their targeted data was Physical data from google fit and academic data exactly like these thesis models. However, their data organization was noticeably clear, but their sample was limited. Although there talked about academic learners, they do not implement their work inside any course platform to show sample data.

2.4 The Benefits of Goal Setting Dashboard

Goal setting helps organize institutional strategies and emphasizes self-managing and self-direction for controlling personal behaviors (Kolb & Boyatzis, 1970). After study considerable research over multiple dashboards and case study over their im-

(16)

units succinctly and straightforwardly. Dashboard converts organizational approach- es into objectives, statistics, targets, and notifications. This performance management system monitors the learning process and activities according to academic performance metrics. Learner’s day to day core performance monitoring and triggering issues according to necessary are principle task of goal dashboard setting (TreeAMS, 2019). Implementation of goal dashboard help, in short, medium, and long terms goal achievements.

Large scale data present in different dimensions to reduce the understanding effort of data before analyzing data inside the goal system. Goal setting dashboard helps to customize data, present all data on a small scale, and drills all details inside it (OntimeBi, 2017). This process helps to improve the decision, performance to makes himself self-sufficient and self-managing.

2.5 Summary

In this chapter, our motive shows in three different stages. First, multiple studies show how vital meta-cognition and self-direction for the learner. Secondly, we try to illustrate learning dashboard themes and limitations. We added self-direction based on one great approach and its fundamental limitations. Also discussed there and our works relations. Finally, we try to illustrate dashboard importance in the goal- oriented self-directed approach.

(17)

3 Goal Dashboard Description

This project connects the quantified self (QS) with learning analytics (LA). Though quantified self (QS) focuses on biological, environmental, physical, and behavioral, we mainly focus on physical behaviors to populate our dashboard with the necessary skills (Swan, 2013). In this project, the Self-direction skills development process holds learner LA, QS data. In this chapter, we try to describe those data sources and their data types briefly.

3.1 SELI Project study content

Learning via computer network shows fundamental importance in 2020. During the pandemic of corona (COVID-19), 1.2 billion learners are out of the classroom. As a result, e-learning, whereby teaching is performing remotely and on the digital platform, dramatically increases. An online platform called BYJU increases enrolling students by almost 200% (World Economic Forum, 2020).

However, ICT implementation in learning already dramatically changes learning and teaching dynamics (Oyelere et al., 2019). SELI project designed a learning platform To "provide the opportunity to create courses for several types of disabilities, considering accessibility standards, interaction between students and stimulating the crea- tivity of tutors and students" (Figure 2).

SELI learning platform design, similar to Moodle, using the blockchain method for the first time in any learning environment (Oyelere et al., 2020). The project promotes different advanced and modern learning techniques, such as game-based learning and puzzle-based learning, both real and virtual life (Oyelere et al., 2017;

Oyelere et al., 2020; Yadav & Oyelere, 2020). This project can be a guideline for future educational applications (Oyelere et al., 2020).

(18)

Figure 2: Architecture of SELI platform (Martins et al., 2020).

3.1.1 What is the SELI project?

Developments in information communication technologies (ICT) change the infra- structure of communities and services. But the real scenario is different. Beneficiary groups of ICTs are mostly in affluent counties. The global divide excludes develop- ing nations, disability groups, and some fundamental factors of life, such as education, economy, and many more; the SELI project is lunch for those disadvantaged groups (Nzeako, 2020; Tomczyk et al., 2019). It is an ecosystem that focuses on digital exclusion and the inaccessibility of education for underprivileged groups. This project aims to ensure active technological involvement and ensure teachers' digital competencies for future challenges (Tomczyk et al., 2020).

(19)

Digital divination (Tomczyk et al., 2019) between countries to countries and society to societies is the sum of complex factors. European people are not shown a significantly different graph during internet use. There is the only difference in the internet using policies, other that very few people in a rural area have low access to technology. Peoples in Latin America use the internet mostly for entertainment purposes, whereas the European people have many different projects and government funds for social, educational, industrial, and academic improvements.

The SELI focuses on establishing an open-source project for Lerner and teachers for ICT cultivation (Oyelere et al., 2019; Tomczyk et al., 2019). Collaboration between Latin, Caribbean, and European union by this European-funded project's main target is to develop technological equilibrium between European, Latin, and Caribbean regions.

3.1.2 How does SELI perform?

Smart Ecosystem for Learning and Inclusion (SELI) is an EU-funded project with targeted audiences from Latin, Caribbean, and European Union, and target groups are the elderly, migrants, displaced youths, physically challenged, deaf and blind (Martins et al., 2019).

(20)

The architecture of the SELI.

Figure 3: The architecture of the SELI project (Martins et al., 2019).

The architecture of the SELI project is different from the traditional LMS (Learning Management System). In the SELI system, the developers follow blockchain infra- structure to inter-connect authentication, course organizer, and user. The system follows a semi-automated authentication approach, so teachers (organizers) have all reserved rights for access control of the illegal users. A registered user can find all their subscribed courses and tasks in an organized way. The system tracks users' data, activity sessions, besides course performance. The system is robust, user friendly, and all performance can do in a real-time manner.

In the SELI platform, there a teacher arranges a course, and anyone can learn from the system (Figure 4a).

(21)

Figure 4a: SELI platform for a tutor (front view).

(Figure 4b, 4c) exhibit after authorizing as a tutor; there is an option to create content and publish it. The system also helps to save the course and get help from existing SELI courses.

Figure 4b: SELI platform for a tutor (Navigation bar).

Besides, an organizer can track the content after publishing, and the dashboard section represents all data. Besides, an organizer can track the content after publishing, and the dashboard section represents all data. A content creator also needs to make a syllabus, required hours to complete the course, and contents lists.

(22)

Figure 4c: SELI platform for a tutor (Instructions).

When students come inside SELI, have the student interface (Figure 5a). A learner can subscribe to one or many courses from the list.

Figure 5a: SELI Student platform.

From the course review, section one can access inside content. Many different courses with different languages run online inside SELI (Figure 5b).

(23)

A user finds out the right course for him by clicking any of the running courses.

When a user clicks on a course can see the total duration of the course, syllabus, table of content along with the course details (Figure 5c)

Figure 5c: Course details.

Clicking on the syllabus or course organizer, a learner can read a detailed syllabus or table of the course contents. Figure 5d demonstrate course content for the Design Smart Learning Environment course.

(24)

3.2 Study data from SELI Platform

The learning model of SELI has lots of study data. For supporting the learning system, our Goal software received learners' fundamentals activities.

 Course Title: The user needs to identify a specific course before reading/planning for any. There we have read all course list a user subscribed inside SELI and lists down in our system.

 Progress: Course progress handles, showing the course percentage of completion. After starting a course, when a user performs to archives lessons, they must need a record of accomplishment for the next progress. We read this progress data from the SELI system, and our system makes a GOAL system against it.

 Duration: Systematic study needs a structured plan, and the Study plan needs to understand the required task. The duration of the course illustrates the minimum required time to complete the course.

 Task/Assessments: Inside a course, there are always have some quiz, assign- ments, and challenges to evaluate students. SELI courses have some assessments task; we read both copulated and incomplete works for our systems.

 Others: Without it, we read a user's information (Name, Email, others) to identify specific users and collect and plan them.

3.3 Google fit health data

Google fit is an ecosystem. It allows users to access their health and wellbeing data, updating the device or sensor to the central storage. A user can use data from any of the machines and upgrade them into new machines as.

3.4 Brief discussion

Google fit APIs provides access to health and wellbeing data to associated google user with specific permissions. Here, the OAuth protocol (IETF Trust, 2010, 2012)

(25)

collecting user consent. An authorized application can then access google store, where google blends data from several applications; around 43 health and available wellbeing apps run with google fitness technologies (Nobakht et al., 2020).

In fitness API, users can access the application without a traditional login system, and they do not even need to remember a username and password. A client can take access token using their google account, which is very common in many applications nowadays (Figure 6).

Figure 6: Sign up with Google.

Google fit performs with the following stages.

Figure 7: Google fitness Platform Overview (GoogleFit, 2020c).

(26)

 Google fit API: There are two different APIs are Android and REST API.

Android API can access by an android mobile user while REST API robust and platform-independent.

The sensor framework consists of data Sources, data type, data points, datasets, and sessions. A dataset holds a collection of data points from a single data source, and data types show collections types and do data delete, aggregate, create operations. A data source is the source of sensor data, which arranges inside storage as data points.

Furthermore, Session data present time intervals of different meta sources.

Different data types inside fitness API are (GoogleFit, 2020a):

 Activity: This data type holds activities, workouts, calories, and sleep data.

The most common data are meditation, sleep, gardening, sports, or running.

The activity type arranges data in real-time. If there are multiple activities are running at the same time, it can divide data into different segments. Then, according to the starting and ends time makes different sessions for them.

 Body data types: The data types have data about body fat, weight, Heart rate, and Height. Here body fat percentage of a user measures against body weight.

 Location: In this data type, google stores the speed of person or vehicles. By measuring the revolve per minute (RPM) of a wheel of any vehicle, it can calculate average speed, top speed, total distance, and time.

 Nutrition Data: Nutrition data is a session type of data where a user might fill fields food type and item or amount of meal to measures total consump- tion. This data type also has Hydration data that measure the amount of water taken by the user in a single drink.

Besides, fitness API has various kinds of data, all of them can access via OAuth verification, but some sensitive data require special permissions rather than standard OAuth authentication. Access this data required an application with OAuth diplomat- ic data verification protocol. In Restricted data authentication protocol, OAuth puts an extra layer of protection, which requests owner permissions and accepts or denied the owner to complete the verification (GoogleFit, 2020b). Data comes into restricted

(27)

cus, Cervical position, Menstruation, Ovulation test, Oxygen saturation, and Vaginal spotting.

3.4.1 Health data collected from Google

To be Self-directed, our aim data from google fitness API is Steps, Sleep, and Calo- ries. We try to track the user's walking distance, sleeping time, and Calories burning in a day. To identify a user and makes a secure search system, we also store the below data:

User personal information: As the process of login is token-based login. Google consent makes a Token for the user that access token, refresh token, and scopes. Decod- ing this access token, we achieve a user's email address, name, Unique data source id, etc. However, we stored email addresses to search specific users, Unique id to store data, and the full user name.

Step data is the real-time movement of a person. Present as a walk in the dashboard.

Sleep data is a little bit different because interval values inside sleep data show sleep types. This data values between 0-112. If someone awake, it means the intervals between 0-29 and 30-109 is for sleep (GoogleFit, 2020d). Besides, 111 is REM sleep, and 112 means awake during the sleep cycle. Calories Data is similar to step data. It calculates calories are burning during activities.

3.5 Data Security and OAuth 2.0

Data security is a heavy mathematical, scientific method for protecting data from unauthorized modification and disclosure (Denning, 1982). This project contains more confidential personal data about learners and teachers. For the self-direction implementation system required healthcare, learning information, healthcare data directly relates to the user's personal Google account, and Learning data with the

(28)

party services, such as Facebook, LinkedIn, Google, GitHub, etc. But for our project only with google, because of google fitness data dependency only on google account.

One-time Authentication (OAuth) is a method implemented by an end-user to protect and access their resources without sharing credentials. Typically, applications require a username password for authentication, but in the case of OAuth, the user needs an access token to access the resource. Unlike the traditional process, it can revoke une- ven users or third-parties by limiting scopes.

OAuth 2.0 (IETF Trust, 2012) implementation is not straightforward; the flow chart can exhibit the real scenario (Figure 8). To understand, OAuth 2.0 is required to real- ize four separate roles inside it.

 Resource owner: Data holder having capable of maintaining access to secure properties. The resource owner usually considers an end-user.

 Resource Server: The protected data hosts invoke or revoke access permis- sions to the protected resource via access tokens.

 Client: Web or mobile application or any entity request to access the protect- ed resource is a client.

 Authorization server: This server generates an access token for client ap- plications depending on the grant type (such as client credentials, authorization code, implicit) (Orange, 2020).

(29)

Figure 8: OAuth 2.0 process overview (IETF Trust, 2012).

The OAuth flow complete in three separate stages. The client application performs one-to-one interaction with the resource owner, authorization server, and resource server.

Stage-01: Client Application request secret resource, justifying valid request re- source owner provide a code (there are some other processes in this project we use codebase grant) in the form of a grant. The code contains the owner's client id, secret, and permitted scopes (Okta, 2020).

Stage-02: Client application passed this authorization grant to the Authorization server, where the authorization server generates an access token according to the provided grant type.

{

"access_token": "ya29. a0AfH6SMDxTcqfP4",

"scope": "https://www.googleapis.com/auth/fitness.activity.read", "token_type": "Bearer",

"expires_in": 3600,

"refresh_token": "1//xx-xy-xz"

(30)

There are some Pros and Cons in OAuth 2.0 (Nebinger, 2020),

 The owner can revoke access to unusual users.

 The access token lifetime can increase with a refresh token

 The owner specifies access to scopes.

 Widespread and continually growing.

 There is a risk of code (grant by author) misuse because of the lack of the correct encryption method.

(31)

4 Goal Dashboard.

Learning data management required cognitive tools like dashboards and visualiza- tions. These tools play a vital role in understanding a visual pattern, identity, and necessary reinforcements to guide a significant decision (Brath & Peters, 2004). Goal system established in a visual dashboard to robust the process of analysis, monitor, plan, and reflection. We designed a goal dashboard prototype after multiple case studies about the learning dashboards (Jivet, 2016; Verbert et al., 2013).

4.1 Architecture of the Goal System

Figure 9: Architecture of this Project.

This process eventually targets to provide equipt software, with LA and QS data sets.

Active research on multiple LA dashboard projects tries to access health and learning data to develop a self-direction process with few excellent features.

(32)

UUID of the learner collected inside goal database simultaneously with users logged email address search users information in the SELI database.

Learning data from the SELI database, fitness data from google store, and user information from users authentication stores inside goal database. Then goal database information collects as in the goal dashboard. Then Analysis, Planning, and Monitor- ing data help to find reflections. There are a few key features that the system contains.

 Web Application: the project provides a web application, which means de- vice independents to operate from mobile devices and computers.

 REST API: data collection from separate REST API, project data presents in another REST API provides independence of data collection modification, addition, and subscription to modifiers.

 Graph Representation: Data presents with bar charts to understand the visu- als difference between plan and performance at a glance.

 Learner Analytics: Project directly connect to the right user databases and gathering all analytics to help analysis, planning, monitoring process.

 Navigate Learner: Software fully responsible for real-time data tracking and showing the necessary steps. Instead, the reflation model clarifies incomplete/

complete tasks to navigate the learner toward the desired goal.

4.2 Data representations

Data presentation emphasizes the process of data analysis. Representation failure affects the conveying data to the reader or reviewer, so a calculative well-justified method should ignite for data presentation (In & Lee, 2017). Visual representation is a cognitive process of data illustration to analyze and makes a decision (Jivet, 2016).

We present our visual with some charts to helps the user sens-making.

4.2.1 Data Collection.

We already described our data source and data types, and now we try to define the

(33)

enough, but we try to illustrate the data collection and preparing process in a standard flowchart (Figure 10).

Figure 10: Data collection stages.

 Read data: Data collection process starting by reading data from the data- base; SELI project data add from their data cloud directly and google fit data from google storage collected via REST API.

 Extract: Informations are in the cloud, or the database does not always struc- ture. There needs modification addition, additions, and substruction before ar- ranging them for analysis. For instant, user collection having information users, their taken courses, and course teacher name. But we need the only course name with the user's information.

 Arrangements: Prepare data for analysis with necessary data, makes an ob- ject, and pass them for illustrating in the chart.

(34)

gress depends on proper data analysis. Visual data presentations directly help target planning by improvising analyzing skills (Leitner et al., 2017).

Figure 11: Data Analysis and Planning Relation.

In this process, we help learners visualize their data on what they are performing in SELI and regular activity.

Figure 12a: Data Planning Model.

The dashboard uses a bar graph with numerical and plain text presentations of activities to help users predict necessary improvements. According to predictions, a learner can plan for learning analytics (LA) (Figure 12a) and quantified self (QS) (Figure

(35)

Figure 12b: Data Planning Model.

4.2.3 Data and Goal

Some of the present generation dashboards used radar or spider graphs to present, which is not suited to our work (Jivet, 2016). Our health data brings some mixed (big and small) data where small data comparison and illustration are hard enough with radar charts. Active research on several visual patterns of charts, we find a bar chart is a suitable option for data presentation. The Bar graph provides a straightforward visualization of data to help evaluate, analyze, and decision-making processes.

(36)

performance in the SELI course or daily activity tracked by google. (ii). Righ bar is a goal set by the learner for their study and health improvements.

Numerical conversion: Each information set describes in numerical form. Although data have multiple forms of arrangements inside storages, all data convert to numerical forms for clear understanding during the analysis and plan-making process. For instance, the dataset's assessment information presents as true/false ( True means task complete and false means incomplete). In this project, we convert data into a number and present it in the same unit and scale as data collected, to reduce extra hassle.

Figure 14: Health data and Goal data Presentation.

Color Presentation: The bar for performance and goal are presented in a different color to create a visual difference at first sight. Besides, the cursor hovers on the bar also presents the bar name, value, and types. This visual data analysis, comparison, and dashboard data reflection determine the necessary re-plan for the system. Ac- cording to requirements, the user reinforces some data to achieve the target.

Other data: In the goal dashboard, activities have timeline data besides other data.

Though there is no deadline in SELI course data, our system generates an auto date to maintain the work pace. Users can set a timeline plan for them, and then the system’s generated dates comparison operation show a reflection.

4.2.4 Reflection model

(37)

disciplinary platform (Govaerts et al., 2012; Keestra, 2017). The present generation research develops some computerized models, which can summarize the whole SDS process. It involves all the efforts between ease of each task to find the outcome of these activities. In this section, we try to present a brief description of our reflection model, its architecture, and its benefits.

Self-reflection can be considered a mirror image of individuals' actions. ICT developments bring some dashboard that runs with the cognitive process, which intercon- nects all supporting entities to help self-controlling, self-managing, and self- awareness as parts of reflections of systematic processing of life. For self-reflection, some of this work compares learning data with previous years' learner data, and some other SDS model compares with standard data (Jivet, 2016; Majumdar et al., 2018).

Our system plan with student prediction and standard data as goal data and compare it with learning data.

Self-reflection model data arrangements, calculation, and presentation process:

(38)

 Different reflection tracks, though there are presents separately for study and health, the operation process is quite similar (Figure: 16).

 From both learning and health, regulation improvises self-direction skills.

Figure 16: Reflection model.

This reflection model is a standard approach chosen for this work entirely new in a feature. Study reflection (Figure: 16) contains four notifications are for four different activity; similarly, health reflection have three tasks. In the process, we try to maintain the same sequence as the data present in the data vs. goal model (Figure: 13, 14).

In reflection, the system suggesting what a user should need to take as their next step to complete the course or tasks. The system also prefers any modification necessary it means any re-plan should need to do the course. Whither, Software congratulates users’ on task completion to motivate for future tasks.

4.3 Frameworks and Technologies

This work's outcome is a single-page web application with React.js, JavaScript, Bootstrap 4, UI design with Material UI, graph with chart.js. For back-end node.js, express.js, google fit rest services, OAuth 2.0 service, and MongoDB database.

 Single Page application: Single page application is an application that does not require a refresh page for each user's action, such as page data loading- pictures, style, and databases. These applications just load for once, and each

(39)

the application is significantly lower. The operation of the web application always divides into very small sub-sections called components. According to necessary, only a small section update/reload, as the entire process is not bothering simultaneously, so the process is first and accurate (Jadhav et al., 2015).

 JavaScript and React.js: JavaScript is one of the most powerful web lan- guages at present works. Around 95% of applications are directly or indirect- ly design with javascript (Luong, 2019). JavaScript developed as a script in 1995, and after the release of the ES6 version, it considers as a programming language. To make a web page dynamic, it has many frameworks and libraries for a client site and server site.

Reactjs is the most popular front-end component-based JavaScript library, used to make an interactive single-page user interface (Aggarwal, 2018). The working interface of Reactjs follows the MVC (Model View Controler) pattern. These three interconnecting interface works like user input received by the controller, managed by models to present the states, as lots of programming languages and their libraries used the same patterns, it's relatively easy to use Reactjs with other frameworks without any interferes.

 Material UI and Bootstrap 4: Material UI is a mobile-based designing lan- guage, which components are isolated. Unlike regular CSS, they inject style according to displaying plans, and they are self-supporting. In material UI, the first code writes for mobile and then modifies components using CSS's media query. Google developed this language to give users realistic experiences with 3D effects, animations, and natural motions (MaterialUI, 2020).

Bootstrap 4 is an open-source framework for any web page; basically, it contains a template (HTML and CSS) (Luong, 2019). Bootstrap provides many styles to make a button, image carousel, navigation, modals, forms, typogra- phy, table, and mobile responsiveness for any front-end developments.

(40)

fronts of visual charts. Under the MIT license, it is the best data visualization library render in HTML canvas.

 Nodejs and Express.js: Node.js is a runtime JavaScript environment written in C++, using Google Chrome’s V8 engine for server-side programming (Liang et al., 2017; Rimal, 2019). It operates asynchronously, which means without waiting for ope operation can process requests. Node.js create servers with the help of express and HTTP module, control users' requests ( receive that and process accordingly), connect them to database or cloud and finally provide a JSON output for render then direct inside an HTML file. In a sim- ple word, node.js creates a non-blocking event executing a platform with an excellent performance guarantee at a maximum speed.

Express is a framework for node.js applications for speeding the development process. It's a middleware group to control all kinds of requests and responses inside applications (Rimal, 2019). All applications might need deployments inside any server. Express is such a framework that is responsible for making a node.js server.

 MongoDB: It is a non-relational cloud platform for data storing, editing, and updating. The data storing process of MongoDB is a schema base, and the data type is BSON (Binary JSON). The database can store any data such as string, number, date, and data that a user only needs to make a schema and define their data to reserve in a specific format. Besides, A user can find all data, specific one data (findOne), sort data (both ascending and descending order) also select particular data from the database.

4.4 Chapter Summary

This chapter actively described the architecture, prototypes, and design of our dashboard. Section 4.1 demonstrates the prominent architecture of this project. In section 4.2, the data presentation process, each stage is discussed separately (section 4.2.1 to section 4.2.4, discussed data collection process, data analysis, and planning process, data presentation against goal and reflection model simultaneously), and in section 4.3 discussed frameworks and technologies used in this project.

(41)

5 Research Methodology

A preliminary or pilot case study works as fundamental for formulating questions or hypothesis testing for any research (Noor, 2008). A questionnaire, survey, and interview conduct for the case study over participants to achieve more information in this research.

5.1 Sample

Analyzing the sample generates a friendly meeting with a few students from the different Universities all over Finland. The meeting has two-part, in pre-meeting all supportive system attachments and presentation delivered. As part of the pre- meeting, we requested all users to download the “google fit application” into their mobile phones and subscribe to the SELI platform courses. Then we request them to make some plans for their study and health data improvement using the goal system.

After delivery, all but 12 participants attending us from different universities were under supervision for three consecutive days.

5.2 Introductory Presentation

The project was presented to all the participants thought out the online presentation on the ZOOM platform. The workflow of the system elaborately described and demonstrated full functionalities. All the participants monitored for three days and always tracked their performances, progress, and initiatives.

5.3 Piloting Study

We had conducted a pilot study to reach our research objectives. For research ques-

(42)

5.4 Interview Data Collection and Analysis

The research questions were divided into multiple sub-questions and added some support questions to make the research accessible and informative. These questions were answered by user opinion, descriptive opinion. In order to answer the opinion- based questions, the audience opinion was taken from five separate phases:

 Strongly Agree

 Agree

 Neutral

 disagree

 Strongly Disagree

Descriptive opinion questions were created to ask 'yes,' 'no' questions, and 'yes' an- swers to additional questions about that.

5.4.1 Analysis question Based on RQ 1:

1. What kind of self-direction supportive system user used before the Goal system?

2. Why did they use it?

As: Education, Health, Other.

1. Self-regulated learning is essential, according to the user?

2. Is health management is equally important?

3. These two: Health, Learning phase of self-direction are helping each other?

5.5 Survey Data Collection and Analysis

The HPLP II survey (Majumdar et al., 2019) was performed to address a few study questions, with each survey question between 1 and 5. The method of addressing both problems provides a full picture of the usefulness of self-direction and its sup-

(43)

We partition health data into two different categories Physical activity and stress management. The physical activity contains walking data and calories burning data, whereas stress management having sleep data.

Physical Activity

1. Check calories burning,

As: Users calories burning tracking process importance 2. Plan calories need to burning

As: Users calories burning planning process importance 3. Total walking steps

As: Users walking tracking importance 4. Your planned steps

As: Users walking planning importance Stress management

1. Sleeping plan

As: Sleep planning importance for stress management 2. Enough sleep

As: Sleep tracking essential for assuring enough sleep 3. Relax time

As: System effects on relaxation time

We had some supportive questions for this research. We present the user's response into a graph and demonstrate data in a percentage manner.

(44)

6 Result and discussion

The project implements inside the Seli platform as a student dashboard. For user in- tegration tests, an interview and a survey generated along with participants. There were 2 different types of questions but are mostly objective types of questions. Be- sides, there are some short and descriptive questions to know the user's thinking's deep level. The questions elaborately describe during the whole process to help users understanding precisely points of the questions.

6.1 Results

Interview organize in zoom platform, user feedback, and supportive question also added with the interviews. But the survey is performed via an online link, and attendees respond according to their time.

6.1.1 Research question 01

Figure 17: Self-directed supportive system users.

There was almost 92 percent, 11 out of 12 participants confirmed they were used several different platforms for their studying, health, and hobbies. As a follow-up question, they were asked, “What type of system did you use? If possible, give names?” among of ‘Yes’ answering participants.

91.67%

8.33%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Use

Have you ever use any self-direction supportive system before the Goal system?

Not sure Yes No

(45)

In answer, all of the participants said different self-directed learning tools name.

Some of the participants had used more than one platform. According to the answer, there is a list of those platforms.

 Duolingo

 Science Zone

 Mimo

 Bookly

 Fit

 Mockup

There were asked another follow-up question, “Please explain the reason behind it.”

Almost all participants used Duolingo for their language learning application. As part of hobbies, a few of them used the Science Zone to learn scientific information.

Some participants mentioned Mimo, a programming learning platform, and Bookly as an e-reader note maker, marker. Also, 4 participants confirmed they were using fitness application for walking, cycling, sleep, calories, and workout tracking.

To answer this research question, we arrange a survey, and there were asked 7 different questions.

 At first, we asked about our application, the value of a calorie visualization process to help the user track daily calories burn during cycling or workout.

(46)

A precise result total of 66.7 percent find helpful, 25 percent find averagely helpful, and 8.3 percent find very much helpful.

 They asked helpfulness of the regular participant’s walking step presentation process in this application.

Figure 19: Walk tracking survey result.

A clear visualization from bar 58.3 percent finds helpful, 33.3 percent find averagely helpful, and 8.3 percent find very helpful.

 Planning from our application for step and calories helped them reach their daily personal calories burning and walking goal.

(47)

In the case of calories burning planning, visualization from bar exhibits 58.3 percent finds our application helpful, 33.3 percent find it very helpful (Figure 20a).

In contrast, 8.3 percent find it not helpful (Figure 20a).

Figure 20b: Survey result on planning for Walk steps.

A clear visualization from bar 58.3 percent finds our application helpful, 33.3 percent find it very helpful during walking planning (Figure 20b). Another 8.3 percent find average helpfully (Figure 20b),

 In this survey, there were two stress management-related questions in these questions; we try to understand users able to track their sleep correctly, make a healthy plan for them, and plan system helps in sound sleep observation.

(48)

Figure 21b: Stress Management: System Effect on Users.

There was some mixed answer to the question “By planning ‘I had’ enough sleep.” Around 75 percent find this application helpful, whereas 16.7 percent marks it as averagely helpful, and 8.3 percent find it not helpful at all (Figure 21a). For another question, “sleep tracking helps in sleep plan,”- around 41.7 percent says averagely helpful, 33 percent says helpful, 16.7 percent says not helpful, and 8.3 percent found not helpful at all (Figure 21b).

 Planning Process helpfulness measure was one of our primary concerns. We try to find the user's opinions on the planning process helpfulness on their relaxation time.

Figure 22: Planning Process Effect on Relaxation time.

Around 58.3 percent found it's helpful, 33.2 percent very helpful, and 8.3 percent

(49)

For this question, we asked the user to subscribe SELI course and continue with it.

But all the users complained there was a problem with getting course-related information; after analysis, a server problem had found. However, they can see their subscribed course not able to see any data. Then we updated the server and demonstrated data from my local machine before performing an interview.

 Do you think self-regulated learning is essential?

Figure23: Importance of self-regulation to the participant.

We had 7 responses for these parts; the other 5 could not add to this part of the interview. Among the participant, 57.1 percent was agreed, 28.57 percent strongly agreed, and 14.28 percent was neutral.

 Do you think health management is equally important?

We get the same results for health management, 57.1 percent was agreed,

29%

57.14%

14.29%

0%

10%

20%

30%

40%

50%

60%

User Responses

Do you think self-regulation learning is essential?

Strongly Agree Agree Nutral Disagree Strongly Disagree

(50)

Figure 24: Importance of health management according to the user's.

 Are they support each other?

From the first two question results, it was almost easy to imagine users have equal priority both for study and health. But we asked them are they find they are supportive of each other, almost 57 percent agreed, and around 33 percent strongly agreed with that (Figure 25).

Figure 25: Importance of Health and Study regulation.

29%

57.14%

14.29%

0%

10%

20%

30%

40%

50%

60%

User Responses

Do you think health management is equally important?

57.00%

33.00%

0.00%

10.00%

20.00%

30.00%

40.00%

50.00%

60.00%

User Responses

Are Health, Learning Regulation support each others?

(51)

 Supportive Questions

We had some supportive research questions. Around 11 participants asked these questions and found the following answer illustrated in the graph.

Figure 26: Supportive Research Questions.

Around 72 percent of the user found the application easy to use, another 28 percent not. Then 63 percent said it’s user friendly, 37 percent said no. Almost 100 percent said the login system was straightforward. For learning and a healthy lifestyle, around 55% found this application helpful, but 45 denied this. Finally, 81 percent believe this application helps in self-direction, and more than 90 percent would be willing to recommend others to use the application.

 User experiences and feedback

The application's general view was liked by around 60 percent, and the user interface was acceptable to 70 percent.

However, there were multiple different feedback from the users—the key

(52)

o Server Student data missing: A small number of learners complained about server issues, which created a problem with their continuous data tracking process from the SELI database.

o The Reflection model was unique: The reflection model designed in this work got acceptance from learners because of real-time sugges- tions and feedbacks.

o Clear data presentation: The hover effect helps the learner to under- stand the dataset type. Attendees found an easily understandable and transparent choice of data illustration.

o Very nice planning model: Data planning model with a numerical model made the planning process quicker and faster.

6.2 Discussion

The main goal of this work was meta-skills developments with the help of technological support. To support this process, we deployed a self-regulation helping hand dashboard at the end of this work. Unlike the traditional self-regulation system, this work's context is not bound to only education but also other learners' regular activities. In this section, we try to discuss outcomes elaborately based on the user's responses to research and its supportive questions.

The answer to three research questions shows a scenario of self-direction skill development and its importance. Research question 01, exhibit students, is very much in- volved with different technologies to support the learning process in their day-to-day life. Most of the time, they use technological support for their language learning plan. The learner did make a goal for vocabulary learning daily. Some of the participants also make the daily goal of walking, sleep, and pushups.

In research question 2, there was a survey to determine user physical activity and stress management tracking and planning from this application importance in participant day to day life. The majority percent of participants were satisfied with the tracking and planning system. Although different fitness applications already have health tracking and goal planning features available. This goal system provides learn-

(53)

ing and health management systems together, which helps the SDS module design process.

Finally, Research question 3: a reflection of this whole process was measured. In this question, we try to analyze the user meta-skills (LA and QS) development to help the self-directed skills development process. Most of the participants find self-direction for health and study equally important, and tracking, the planning process develops standard learning skills

6.3 Limitations

We acknowledge some critical limitations of this works. Data source architecture and data projections create some fundamental problems are below. Mobile devices have no sensor to read sleep data, but wearable devices (e.g., smartwatch, band) have those sensors. So wearable devices only can record the participant's sleep data; those who do not have a wearable device will miss sleep data. Besides, Google fit data can only access a Google account, so the user must need a Gmail account to use this application.

Considering the necessity of data, we have 3 different health-related data in this project to maintain learner activity and stress. But some other factors should need to include for batter performance. Finally, The system works only with the SELI platform learning data.

Self-Regulated Learning Platform Based on Learning and Health Data