Research Project Supervision¶
- PhD or Master Supervision
- A Platform for Air Pollution Detection NEW
- Machine Learning Algorithms to Detect Learning Tactics
- Analysis of Behavioral Patterns after Data Visualization
- Goal Setting Application for Self-Regulation in a User Experience
- Back-annotation of Web Resources with Analytics Visualizations
- Visualisation of Discussion Forum Activity
- A Platform to Create Interactive Learning Material using Markup Language
- Finished Projects
I currently supervise the following types of projects:
- Undergraduate thesis projects (UG)
- Coursework projects (CW)
- Master by coursework programs (MC)
- Master by research programs (MR)
- PhD (PH)
For the last two, Master by research and PhD, you may find similar information in Sydney Research Supervisor Connect.
The projects are in the areas of behavioral analytics and data mining and machine learning algorithms. These areas cover the use of data mining and analytics methods to detect and support users, and in general how to integrate the use of technological platforms as part of the design of a user experience. Topics include data collection in interactive environments for analytics, data visualization, data-based personalization, algorithms for recommendation, applied machine learning algorithms to detect behavior patterns, collaborative work analysis, video analytics, content creation for data collection and support, etc.
- Proficiency in web application development (Django, NodeJS preferred)
- Proficiency in statistical analysis (use of tools such as R or SPSS)
- Outstanding communication skills (no, seriously!)
- Experience in Data visualization
Aside from the assessment required in the Capstone project units (ELEC4720/1, ELEC5020/1, ELEC55520/1), I require students to:
this topic description templateto write the topic description document.
- Manage the code and documents in the project through either github or bitbucket repository.
- Work regularly in the project and reflect this work (whatever it is) in the bitbucket/github repository. Your sustained effort will be assessed looking at the logs in the shared repository.
- Write a two paragraph document and send it to me by COB every Friday during the project reporting what was done during the week and what is planned for next week.
- Use a wiki page to write what your project is about and showcase it with a video. You may use the wiki for additional tasks such as document storage, resource collection, manual for your product, etc.
PhD or Master Supervision¶
When searching for Master or PhD supervision, please send an email to me with the following documents:
- A copy of your transcript
- A short proposal of your project idea. I expect 2-3 pages written in proper academic English, with bibliographical references. It should address a topic that is relevant and propose novel directions to explore. I suggest you review my latest scientific publications.
If you are an international student, you will need to find your own funding, that is living expenses and tuition fees. Please visit the Graduate School of Engineering for information about costs and scholarships.
If the initial report and qualifications are appropriate, I will ask you to write a more concrete proposal and potentially schedule a phone conversation. PhD programs in Australia are 3-4 year long. Masters are 1-2 years.
These are the descriptions of the projects currently available. Make sure you check that the level is aligned with what you are looking for.
A Platform for Air Pollution Detection NEW¶
The current graduate attributes expected from a graduate in Engineering is the capacity to work in multidisciplinary projects that combine areas of expertise in which engineering solutions need to be adapted to another domain of knowledge.
The project is proposed in the context of environmental toxicology. Air pollution has been identified as one of the top causes of premature death in the world after high blood pressure, smoking, high blood sugar and cholesterol. The International Agency for Research on Cancer (IARC) has classified air pollution as a group 1 carcinogen and leading cause of environmental cancer deaths. As a consequence, there is a pressing need to design systems that can detect certain toxic dust particles in real time and translate these readings into indicators of air quality. These systems would be deployed in monitoring stations in highly populated areas or those prone to poor environmental conditions to alert authorities when the levels of pollution reach certain limits.
The project needs to combine the two areas and therefore is expected to be executed by one engineering student and one toxicology/pharmacology student in close collaboration. The main requirements for the project in these two areas are:
- Engineering: Design and implement the system that performs regular readings of a dust sensor and sends them to a platform in which these readings are archived and then retrieved either through visualisations, or through reports. The system has to assume that the sensor is read with certain frequency and the response to certain readings may need to be almost immediate (certain reasonable delay can be considered). Additionally, the subsystem in charge of archiving the measurements must assume that there may be a network of sensors providing information from different locations and take that into account when providing reports and visualisations.
- Pharmacology: Assess the risk and identify the hazard that needs to be identified: the size of particles that need to be detected. Provide the rationale for the experiment (to guide the engineering design) and establish the requirements for measurement and detection of changes in exposure over time.
Machine Learning Algorithms to Detect Learning Tactics¶
Description: Electronic tools that support learners provide a wealth of data about the interactions occurring in a learning environment. This data can be analyzed using machine learning algorithms to detect patterns that suggest learning tactics. The detection and analysis of these patterns offer additional insight on the learning process and provide opportunities for the creation of personalized learning environments.
Challenge: Connect the results derived from the machine learning algorithms with suggestions to users and verify that these suggestions are useful.
Level: MC, MR, PH (See project classification)
Analysis of Behavioral Patterns after Data Visualization¶
Description: Dashboards offer users visual information about a specific context. But, what is the effect of these visualizations? Do they prompt any change in user behavior? Are they only a collection of neat graphics that catch the attention of users only for a brief moment? Research is needed to measure the impact of dashboards in users. The project proposes the use of data mining and temporal analysis to detect and characterize the usage patterns before and after a visualization has been seen by a user.
Challenge: Obtain a detailed characterization of user behavior before and after accessing the visualization. Characterize the differences in behavior with robust statistical methods.
Level: MC, MR, PH (See project classification)
Goal Setting Application for Self-Regulation in a User Experience¶
Description: User behavior can be influenced by proposing a goal setting cycle in which a set of goals is proposed, users choose those that are feasible, are given information about their progress, and prompted for reflection at the end of the process. The project consists on defining this cycle as a tight loop and provide a web-based application to support the entire experience.
Challenge: Connect this application with a real-life scenario in which goal-setting can make a difference. Shape the application and include a set of pre-identified desirable outcomes, a set of indicators about the progress, and a set of reflection techniques.
Level: MR, PH (See project classification)
Back-annotation of Web Resources with Analytics Visualizations¶
Description: The interactions with resources while users participate in an interactive experience can be collected and used to produce visualisations to support reflection. Reflection is a powerful technique to increase self-regulation which in turn is shown to promote higher engagement. However, most of the information captured during these experiences is either not properly integrated so it can be given to the users, or it can be accessed through a convoluted process rendering it not reachable. This project proposes the implementation of a back-annotation system by which the data collected is integrated as part of the available resources, that is next to the relevant material. For example, if the interaction includes a video, and there is information about the type of engagement with that video, the data visualization should be part of the activity and available embedded in the same page as that resource.
Challenge: Create a web-application that provides a single point of visualization for both the material and the visualizations. Ideally, a workflow including content creation should be proposed. The project can reuse two already existing platforms to a) collect the data and b) create the material.s
Level: UG, CW, MC, MR (See project classification)
Visualisation of Discussion Forum Activity¶
Description: Discussion forums are one of the main building blocks of a large percentage of learning scenarios. Instructors need to observe the interactions occurring in a forum, assess their quality, and even propose actions to improve the communication. The project proposes the creation of intuitive visualisations of discussion fora with large number of students. The resulting graphs should offer instructors information about interaction, centrality, density, etc.
Challenge: Design a set of graphic visualisations that capture and convey in an intuitive way the structure of a discussion forum to an instructor.
Level: CW, MC, UG (See project classification)
A Platform to Create Interactive Learning Material using Markup Language¶
Description: Creating electronic learning material for courses is still one of the most complex tasks surrounding a learning scenario. With the inclusion of aspects such as student created content, active learning, and continuous evaluation, content is no longer a static entity that is created at some point and left untouched during a learning experience. Instead, new workflows are needed that blend aspects such as dynamic content creation, moderation, analytics, etc. The project proposes to use RestructuredText as the source language to describe the content and the platform Sphinx to translate the content to HTML5. The contribution is the definition of a workflow that allows a set of users to collectively design an publish almost in real time content for a learning environment.
Challenge: Include in the workflow support for operations such as merging contributions, differentiated roles (designer, writer, manager, etc.), and flexible publishing procedures (ssh, ftp, webdav, etc.).
Level: UG, CW, MC (See project classification)
- Pardo, A. (2014). Reauthoring: A Toolkit to create Liquid HTML Active Learning Resources. Available in https://bitbucket.org/abelardopardo/reauthoring..
- Pardo, A., Fisteus, J. A., & Delgado Kloos, C. (2012). A Distributed Collaborative System for Flexible Learning Content Production and Management. Journal of Research and Practice in Information Technology, 44(2), 203–221..
These projects were proposed and executed in the previous semesters. You may use them as a reference.
Automatic Testing of Arduino Programs¶
Description: Arduino is a computer system containing a microcontroller of the AVR family that is used for tasks involving sensors and actuators. There is an enormous community of users of this system and plenty of tools to program and create cool applications. The objective of the project is to put this platform in an education scenario. Imagine thousands of people trying to program a given task in Arduino. How do you test it? The challenge is that to test a program you need a board and sometimes additional hardware. Can you design a program that takes an Arduino program and tests a few features? Can you emulate the required hardware with software? Can you detect and verify the patterns coming out of the output pins?
Challenge: Create a versatile software tester for non-expert users. In other words, if a user creates an Arduino program, it should be possible to specify a few tests that the emulator would take and then apply to a large number of identical programs.
Level: UG, CW, MC (See project classification)
Automatic Collaborative Spaces Management in MoinMoin¶
Description: Collaborative spaces are very useful to support numerous pedagogical activities. But the advantages fade quickly when scaling them to large number of users. Handling simple operations such as addition, removal or re-configuration of spaces for large numbers of teams becomes too complex in conventional platforms. The project proposes the automation of these management tasks. A platform is needed to support the creation and management of collaborative spaces in the wiki platform MoinMoin. The application needs to support assignment of students to teams, reassignments, permission management, creation and propagation of templates to spaces, bulk processing of documents in all spaces, etc.
Challenge: Create a program that can be easily used by instructors to handle complex operations over teams and see the effect in a moinmoin wiki almost instantaneously. User trials are highly recommended.
Level: UG, CW (See project classification)
Design of Student Dashboards for Learning Analytics¶
Description: A dashboard is a collection of data visualizations that helps users understand complex data obtained from a particular environment. In the context of learning, a dashboard is useful to show students and instructors indicators of certain activities. The project consists on designing a dashboard that is intuitive and useful for students and instructors.
Challenge: Create a web application with the dashboard that reads a database with information and creates the visualizations for a large population of students and instructors. The application has to be validated with user trials.
Level: UG, CW, MC (See project classification)