Available Workshops & Tutorials

We are pleased to present a great LASI’17 program that includes 7 Workshops and 11 Tutorials!

Individuals take a deep dive into one workshop across the entire LASI, while tutorials enable participants to get a flavour of a range of topics.

Each individual can choose 1 workshop and 2 tutorials.

WorkshopsTutorials
W1. Building predictive models of student success with the weka toolkit
Title: Building predictive models of student success with the weka toolkit

Chris Brooks, University of Michigan
Craig Thompson, University of Saskatchewan

Description: In this workshop participants will learn the theory behind several computational methods for machine learning and predictive modelling, including decision trees, naive bayes, k-means clustering and expectation maximization clustering. Participants will be able to use these methods on a supplied set of data using the freely available weka toolkit. The final session of the workshop will be dedicated to exploring new data (and participants are encouraged to bring datasets they are interested in with them).

Prerequisite Skills/Experience: Participants should be familiar with basic introductory statistics. No programming experience is necessary. This workshop is particularly aimed at individuals who do not have experience with predictive modeling techniques.

Advanced preperation:

  • Participants should download and install the latest version of Java for their operating system, as well as the Weka toolkit (version 3.8, http://www.cs.waikato.ac.nz/ml/weka/downloading.html).
W2. SNA for Learning Analytics in Formal and Informal Learning Environments
Title: SNA for Learning Analytics in Formal and Informal Learning Environments

Sean Goggins, University of Missouri

Description: Course management systems, games for learning, synchronous collaboration systems and other technologies each have different electronic trace data technical characteristics. Social Network Analysis (SNA) is a research method applied to the analysis electronic trace data (log files) from these environments. Typically, SNA for Learning Analytics is used in concert with complementary research methods, including computational linguistics, ethnography and ethnomethodologically informed analysis of participant behavior.

Participants will work through examples from learning management systems, collaborative software construction and games for learning environments. Participants are also encouraged to submit their own data for pre-workshop assessment and discussion.

The SNA approach to Learning Analytics focuses on the Group Informatics Model and its related, two-phase methodological approach in detail. Group Informatics is the workshop organizer’s systematic methodology and ontology for scaffolding researcher decision making. Phase one of the methodological approach centers on a set of guiding research questions aimed at directing the application of Group Informatics to new corpora of integrated electronic trace data and qualitative research data. Phase two of the methodological approach is a systematic set of steps for transforming electronic trace data into weighted social networks. We show that the Group Informatics methodological approach is the starting point for important discussions aimed at advancing empirically and theoretically informed analysis of electronic trace data focused on small groups. Group Informatics can also be used as a foundation for pursuing research questions in a range of technology mediated environments where formal and informal learning take place. Participants will learn how to perform SNA technologically; but the frame of that technology use for research is much broader.

Prerequisite Skills/Experience:

  • A nominal list of 4-9 readings. At least four will be required, but they are conceptual in nature.
  • Some familiarity with R or Python programming will be helpful, but is not required.

Advance Preparation:

  • See the workshop Website: http://www.sociallycompute.io
W3. Automated Personalization in Education
Title: Automated Personalization in Education

Charles Lang, Teachers College

Description: Automated personalization of student interventions has been a long-standing goal of the educational enterprise. From Skinner’s “teaching machines” to Khan Academy, the idea that technology can use individual student characteristics to ensure the most appropriate pedagogical approach is applied at the most advantageous time has inspired generations of educationalists and inventors. This workshop will demonstrate a workflow for devising automated personalization strategies. It will cover basic concepts in personalization (differentiation, individualization, etc), describing educational goals, analytic strategies for identifying points of leverage, statistics for validating these points, and monitoring systems for fidelity and bias. Participants will develop a project based on a personalization strategy of their own interest.

Prerequisite Skills/Experience:

  • None, but advanced preparation highly recommended. Workshop instruction will be conducted in R.

Advance Preparation:

  • Minimum for participation: Follow installation steps at this address: https://github.com/charles-lang/LASI17-Personalization/blob/master/README.md
W4. Writing Analytics
Title: Writing Analytics

Andrew Gibson, University of Technology Sydney
Simon Buckingham Shum, University of Technology Sydney

Description: A large majority of academic disciplines focus on the development of learners’ skills in critical review, conceptual synthesis, reasoning, and disciplinary/professional reflection. In these subjects, writing arguably serves as the primary window into the mind of the learner. Beyond scholarly academic writing, there is also interest in disciplined, autobiographical reflective writing as a way for students to review and consolidate their learning, thus providing a means for assessing the deepest kinds of shifts that can occur in learner agency and epistemology. However, writing is time consuming, labour-intensive to assess, difficult for students to learn, and not something that all educators can coach well, or even consider their job to coach.

Writing Analytics aims to address these challenges through computational analysis with Natural Language Processing (NLP) techniques. In a pedagogical context, Writing Analytics provides opportunities for scalable, near real-time feedback to learners and teachers, for the purposes of assisting learners improve the quality of their writing. In our view the goal is to move beyond the assessment of texts divorced from contexts, transitioning instead to a more nuanced investigation of how analytics may be effectively deployed for learning.

This workshop will introduce participants to some of these broader issues, before getting hands-on with relevant NLP tools in order to obtain a deeper understanding of the potential and pitfalls. Throughout the workshop we will engage with the following questions:

  • How do we connect low level text analysis features (e.g. word vectors, topic models, parts of speech, lexical metrics), to higher level constructs of a learner’s writing?
  • What are the various assumptions underpinning the computational models we use?
  • How can we take a pedagogically informed perspective when working with computational models and algorithms?
  • How do we measure the efficacy of our writing analytics approaches?

Prerequisite Skills/Experience:

  • Workshop participants should have at least a basic level of programming experience and at least a familiarity with NLP tools for their language of choice (e.g. Stanford CoreNLP and Apache OpenNLP for JVM languages, NLTK for Python).
  • We will use a Jupyter style Notebook for some of the hands on exercises, but more experienced participants may prefer to try examples in their language of choice.

Advance Preparation:

You will need at least one example of student writing to work with for the workshop. More than three may be useful and necessary if the text is 1 page or less.
Prior to the workshop, try the following tasks:

  1. Clean and tokenise your texts
  2. POS tag your texts, and create Maps (java/scala) or dicts (python) of words and their postags.
  3. (Optional) Build a topic model using an algorithm like LDA and assign topics to each of your texts (or to each paragraph of a single text).

For the broader context, and examples of different approaches, browse the motivations and materials from recent LAK Writing Analytics Workshops:

To learn more about the UTS:CIC Writing Analytics approach, and how we seek to ground it in the pedagogy of writing:

  • Gibson, A., Aitken, A., Sándor, Á., Buckingham Shum, S., Tsingos-Lucas, C. and Knight, S. (2017). Reflective Writing Analytics for Actionable Feedback. Proceedings of LAK17: 7th International Conference on Learning Analytics & Knowledge, March 13-17, 2017, Vancouver, BC, Canada. (ACM Press). DOI: http://dx.doi.org/10.1145/3027385.3027436. [Preprint] Awarded Best Full Paper.
  • Knight, S., Buckingham Shum, S., Ryan, P., Sándor, Á., & Wang, X. (In Press). Designing Academic Writing Analytics for Civil Law Student Self-Assessment. International Journal of Artificial Intelligence in Education, (Special Issue on Multidisciplinary Approaches to Reading and Writing Integrated with Disciplinary Education, Eds. D. McNamara, S. Muresan, R. Passonneau & D. Perin). Open Access Reprint: http://dx.doi.org/doi:10.1007/s40593-016-0121-0
W5. From Data to Student Support Actions
Title: From Data to Student Support Actions

Abelardo Pardo, University of Sydney

Description: Educational institutions are dealing with the paradox of having access to an unprecedented amount of data about the student learning experience, and a lack of the skills and processes to transform this data into knowledge that support decision making at different levels. Although other disciplines such as finance, marketing, health or sports have harnessed the power of data to obtain quantifiable improvements in their context, these improvements in education seem to be elusive.

Educational institutions are usually highly complex institutions with organisational units supporting different aspects of the overall student experience. However, the data required to support students is usually scattered among these institutions and requires non-trivial processes and skills to deploy the required know-how to distil knowledge, and ultimately, use the knowledge to support decision-making procedures. Understanding this complex landscape requires being aware of the building blocks articulating the connection between primary data sources (enrolment data, events recorded by the learning management system, etc.) and the provision of personalised student support actions. This connection involves elements such as data collection points, data management infrastructure, data analysis, data-sensitive design, data-driven support programs, policies to sustain these elements, and a framework that promotes continuous evaluation and improvement.

The workshop will expose attendees to the elements required to establish this connection starting with the collection of data at the institutional level and following this journey until the data is used to support the goals of a learning design through the deployment of personalised support actions.

At the end of the workshop, attendees should be able to:

  • Describe the required steps to extract data from various platforms within an educational institution and prepare it for its use in teaching and learning contexts.
  • Demonstrate the use of basic data manipulation procedures to obtain logs, create basic visualisations, and use predictive models while supporting students.
  • Show a sample of the tools and processes required to derive knowledge from data collected during a learning experience.

Prerequisite Skills/Experience:

  • Basic familiarity institutionally-held student data.

Advance Preparation:

Participants should bring their own computer to the workshop with the following tools installed:

  • Microsoft Excel
  • Python (Programming Language + a visual editor + its dependencies)
  • Rstudio (and its dependencies such as the R package)

Optionally, attendees should bring to the workshop a sample of the type of data that is available on their institution.

W6. Multi-Modal Analytics
Title: Multi-Modal Analytics

Xavier Ochoa, Escuela Superior Politécnica del Litoral

Description: Learning does not only occur over Learning Management Systems or digital tools. It tends to happen in several face-to-face, hands-on, unbounded and analog learning settings such as classrooms and labs. Multimodal Learning Analytics (MMLA) emphasizes the analysis of natural rich modalities of communication during situated learning activities. This includes students’ speech, writing, and nonverbal interaction (e.g., movements, gestures, facial expressions, gaze, biometrics, etc.). A primary objective of multimodal learning analytics is to analyze coherent signal, activity, and lexical patterns to understand the learning process and provide feedback to its participants in order to improve the learning experience. This workshop is posed as a gentle introduction to this new approach to Learning Analytics: its promises, its challenges, its tools and methodologies. To follow the same spirit of MMLA, this workshop will include a hands-on learning experience analyzing different types of signals captured from real environments.

Prerequisite Skills/Experience:

  • Basic knowledge of Learning Analytics
  • Programming in any language
  • Basic Machine Learning knowledge

Advance Preparation:

  • The participants should be required to install a virtual machine in their laptops in order to perform the different activities during the workshop.
W7. Analyzing Temporal Data: Theory & Tools
Title: Analyzing Temporal Data: Theory & Tools

Britte Haugan Cheng, SRI
Erica Snow, SRI

Description: Temporal analytics explore temporal aspects of learning and teaching, to gain insights into the processes through which learning occurs. The study of temporal patterns in learner data (e.g., changes in how students access web resources over time, interact with peers) and their relationship to learning outcomes is a central area of interest. This workshop will present a set of approaches and tools to conduct temporal analytics, including hands-on sessions to apply those approaches with a common data set.

Using multiple lenses to analyze a common data set will ground workshop discussions of the theoretical and practical issues of various temporal approaches, constructs and patterns that are visible via these approaches. We will also consider issues of interpretation and visualization of results that are critical to all learning analytics but are particularly challenging for temporal analytics due to their complex and dynamic character.

Prerequisite Skills/Experience:

  • None

Advance Preparation:

  • Participants should bring their own computer to the workshop. Links will be provided to participants to download any data sets and tools that will be used during the workshop. Tools will most likely include R and RStudio, Tableau, and The Ribbon Tool.
T1: Building predictive models of student success with the weka toolkit
Title: Building predictive models of student success with the weka toolkit

Chris Brooks, University of Michigan
Craig Thompson, University of Saskatchewan

Description: In this tutorial participants will learn the theory behind several computational methods for machine learning and predictive modelling, including decision trees, naive bayes, k-means clustering and expectation maximization clustering.

T2: SNA for Learning Analytics in Formal and Informal Learning Environments
Title: SNA for Learning Analytics in Formal and Informal Learning Environments

Sean Goggins, University of Missouri

Description: Course management systems, games for learning, synchronous collaboration systems and other technologies each have different electronic trace data technical characteristics. Social Network Analysis (SNA) is a research method applied to the analysis electronic trace data (log files) from these environments. Typically, SNA for Learning Analytics is used in concert with complementary research methods, including computational linguistics, ethnography and ethnomethodologically informed analysis of participant behavior.

T3: Automated Personalization in Education
Title: Automated Personalization in Education

Charles Lang, Teachers College

Description: Automated personalization of student interventions has been a long-standing goal of the educational enterprise. From Skinner’s “teaching machines” to Khan Academy, the idea that technology can use individual student characteristics to ensure the most appropriate pedagogical approach is applied at the most advantageous time has inspired generations of educationalists and inventors. This tutorial will discuss a workflow for devising automated personalization strategies. It will cover basic concepts in personalization (differentiation, individualization, etc), describing educational goals, analytic strategies for identifying points of leverage, statistics for validating these points, and monitoring systems for fidelity and bias.

T4: Writing Analytics
Title: Writing Analytics

Andrew Gibson, University of Technology Sydney
Simon Buckingham Shum, University of Technology Sydney

Description: Writing Analytics sits at the convergence of Natural Language Processing (NLP) and the pedagogy of writing. It involves the computational analysis of writing in order, ultimately, to improve it. In our view the goal is to move beyond the assessment of texts divorced from contexts, transitioning instead to a more nuanced investigation of how analytics may be effectively deployed in different contexts. This tutorial will introduce participants to some of these broader issues,and get an overview of the relevant NLP tools as way to get a deeper understanding of the potential and pitfalls.

T5: From Data to Student Support Actions
Title: From Data to Student Support Actions

Abelardo Pardo, University of Sydney

Description: Educational institutions are dealing with the paradox of having access to an unprecedented amount of data about the student learning experience, and a lack of the skills and processes to transform this data into knowledge that support decision making at different levels. Although other disciplines such as finance, marketing, health or sports have harnessed the power of data to obtain quantifiable improvements in their context, these improvements in education seem to be elusive.

Educational institutions are usually highly complex institutions with organisational units supporting different aspects of the overall student experience. However, the data required to support students is usually scattered among these institutions and requires non-trivial processes and skills to deploy the required know-how to distil knowledge, and ultimately, use the knowledge to support decision-making procedures. Understanding this complex landscape requires being aware of the building blocks articulating the connection between primary data sources (enrolment data, events recorded by the learning management system, etc.) and the provision of personalised student support actions. In this tutorial, students will learn about the connections between elements such as data collection points, data management infrastructure, data analysis, data-sensitive design, data-driven support programs, policies to sustain these elements, and a framework that promotes continuous evaluation and improvement.

T6: Multi-Modal Analytics
Title: Multi-Modal Analytics

Xavier Ochoa, Escuela Superior Politécnica del Litoral

Description: Learning does not only occur over Learning Management Systems or digital tools. It tends to happen in several face-to-face, hands-on, unbounded and analog learning settings such as classrooms and labs. Multimodal Learning Analytics (MMLA) emphasizes the analysis of natural rich modalities of communication during situated learning activities. This includes students’ speech, writing, and nonverbal interaction (e.g., movements, gestures, facial expressions, gaze, biometrics, etc.). A primary objective of multimodal learning analytics is to analyze coherent signal, activity, and lexical patterns to understand the learning process and provide feedback to its participants in order to improve the learning experience. This workshop is posed as a gentle introduction to this new approach to Learning Analytics: its promises, its challenges, its tools and methodologies. To follow the same spirit of MMLA, this workshop will include a hands-on learning experience analyzing different types of signals captured from real environments.

T7: Learning analytics Community of Practice, human power driving data discovery at the university scale
Title: Learning analytics Community of Practice, human power driving data discovery at the university scale

Pablo Munguia, Royal Melbourne Institute of Technology University (RMIT)

Description: This dynamic two hour tutorial will showcase the benefits of a learning analytics CoP through active participation. This CoP model relies on a team-based learning approach producing deliverables that service both academics and university-wide administration.

T8: Black Box Learning Analytics? Beyond Algorithmic Transparency
Title: Black Box Learning Analytics? Beyond Algorithmic Transparency

Simon Buckingham Shum, University of Technology Sydney

Description: As algorithms pervade societal life, they’re moving from an arcane topic reserved for computer scientists and mathematicians, to the object of far wider academic and mainstream media attention (try a web news search on algorithms, and then add ethics). As agencies delegate machines with increasing powers to make judgements about complex human qualities such as ’employability’, ‘credit worthiness’, or ‘likelihood of committing a crime’, we are confronted by the challenge of “governing algorithms”, lest they turn into Weapons of Math Destruction. But in what senses are they opaque, and to whom? And what is meant by “accountable”?

The education sector is clearly not immune from these questions, and it falls to the Learning Analytics community to convene a vigorous debate, and devise good responses. In this tutorial, I’ll set the scene, and then propose a set of lenses that we can bring to bear on a learning analytics infrastructure, to identify some of the meanings that “accountability” might have. It turns out that algorithmic transparency and accountability may be the wrong focus — or rather, just one piece of the jigsaw. Intriguingly, even if you can look inside the algorithmic ‘black box’, which is imagined to lie in the system’s code, there may be little of use there. I propose that a human-centred informatics approach offers a more holistic framing, where the aggregate quality we are after might be termed Analytic System Integrity. I’ll work through a couple of examples as a form of ‘audit’, to show where one can identify weaknesses and opportunities, and consider the implications for how we conceive and design learning analytics that are responsive to the questions that society will rightly be asking.

T9: Developing and implementing an institutional data governance policy
Title:Developing and implementing an institutional data governance policy

Grace Lynch, Royal Melbourne Institute of Technology University (RMIT)
Stephanie Teasley, University of Michigan

Description: Data governance is a set of roles and policies that are combined to improve how the data assets are handled within an organisation. The policy should establish a set of rules defining how decisions about data are made, who is accountable for its management, who can access it, and how data is extracted, obtained and stored. This policy needs to be integrated with current organisational processes and practices in order to translate into tangible improvements in how data assets are managed and used across the institution.

The interdependency between institutional data and its business processes and applications increases the need for clear guidelines for regarding how such data is managed.This tutorial will discuss how to:

  • define roles and responsibilities in relation to the governance of institutional data;
  • identify the best practices in data management to facilitate its use within the institution, provide a secure
  • environment for its access and analysis whilst ensuring the privacy of all stakeholders;
  • define the organisational structure of roles, access rights and responsibilities;
  • establish clear lines of accountability; and
  • assure that the University complies with the current laws, regulations and standards about data management.
T10: Statistical Analysis of Interdependent Observations in Learning Environments: Exponential Random Graph Modelling (ERGM)
Title: Statistical Analysis of Interdependent Observations in Learning Environments: Exponential Random Graph Modelling (ERGM)

Nia Dowell, University of Michigan
Oleksandra Skrypnyk, University of South Australia

Description: Learning analytics researchers apply statistical analysis to the data collected from educational settings to quantify learning processes. The data collected in such environments often contain interdependent observations. For instance, when a learner chooses to interact with one learning resource, she by necessity ignores other options, therefore, the resources selected are not independent of one another. Similarly, when students collaborate, their interactions with one another cannot be seen as independent observations. Such data is inherently relational, and cannot be analysed with many conventional statistical techniques that assume independence.

Statistical techniques developed in network science aid in understanding and quantifying patterns found within such relational datasets. Network researchers have developed techniques for relational data that include block modelling, relational event modelling, exponential random graph modelling and stochastic actor-oriented models, among others. These techniques are relatively new, and their application in the learning sciences has so far been scarce. Given that relational data is common in learning environments, an understanding of these techniques could be of great relevance to the learning analytics community.

This workshop will introduce one such method of relational analysis, namely exponential random graph modelling (ERGM). ERGM has been applied to self-reported student networks both at the class and school levels (e.g., Heidler et al., 2014; Koskinen & Stenberg, 2012), and recently to the communication networks in MOOCs (Joksimovic et al., 2016; Kellogg et al., 2014; Poquet, Dawson, Dowell, 2017). As such, ERGM belongs to a family of probability models that estimate the probability of network ties to occur, given that certain local social processes and factors external to the network, such as actor characteristics, can constrain tie formation (Robins et al., 2007; Lusher, Koskinen & Robins, 2012). ERGM necessitates a series of theoretical considerations for its model construction, and involves rigorous checks of model fit.

The aim of this workshop is to introduce the basics of ERGM, and demonstrate their application to learning analytics. Upon the completion of this 2 hour tutorial, the participants will:

  1. understand how to evaluate an ERGM analysis presented in a paper
  2. know the steps involved in ERGM analysis
  3. have sufficient resources to get started with one’s own ERGM analysis

Prior experience in programming is not required. However, a basic understanding of social networks and social network analysis is necessary.

Resources for this tutorial will be made available here: https://github.com/ndowell/LASI17-ERGM-Tutorial

T11: Analyzing Temporal Data: Theory & Tools
Title: Analyzing Temporal Data: Theory & Tools

Britte Haugan Cheng, SRI
Erica Snow, SRI

Description: Temporal analytics explore temporal aspects of learning and teaching, to gain insights into the processes through which learning occurs. The study of temporal patterns in learner data (e.g., changes in how students access web resources over time, interact with peers) and their relationship to learning outcomes is a central area of interest. This workshop will present a set of approaches and tools to conduct temporal analytics, including hands-on sessions to apply those approaches with a common data set.

Using multiple lenses to analyze a common data set will ground workshop discussions of the theoretical and practical issues of various temporal approaches, constructs and patterns that are visible via these approaches. We will also consider issues of interpretation and visualization of results that are critical to all learning analytics but are particularly challenging for temporal analytics due to their complex and dynamic character.