MSc and UG Project Ideas

Chris Greenhalgh, 2017-05-18

Preferred Projects

These specific projects are all suitable for HCI or CS

Muzicodes

The muzicodes system (http://dl.acm.org/citation.cfm?id=2986444) allows musical phrases to be used as triggers in an interactive system. For example, when a musician plays the first few notes of a tune this might be defined as a code that starts the accompaniment. There are several possible projects that would use and build upon this system (and others would also be welcome).

M1. Muzicodes Online

Design, prototype and evaluate a social online version of the muzicodes system. For example, this could be a competitive platform where one user can set musical challenges for other users. Or it could be a collaborative/user-generated content platform, where users share compositions incorporating muzicodes.

M2. Muzicodes Ensemble Performance

Design, if necessary extend and then evaluate an application of muzicodes to live group (rather than individual) performance. You would need to identify and recruit the group of musicians to work with. You would then need to work with them to explore and work up one or more specifically collaborative ways of using muzicodes within a performance setting. Given the dynamic nature of muzicodes the performance would need to have a significant element of performer choice, e.g. semi-improvised elements.

Databox

Databox is a platform that allows individuals and households to manage their personal data and allow other parties to run apps or computations against their data without giving their data away (http://www.databoxproject.uk/ https://youtu.be/NBxMk4LMGbk?list=PL9UQcjjmGF3cUoLmvvfrepPAM62sxx5_3 http://www.eecs.qmul.ac.uk/~hamed/papers/CAN2016Databox.pdf http://www.eecs.qmul.ac.uk/~hamed/papers/Human-Data-Interaction.pdf). The first public developer release of Databox will be in March 2017. The kinds of data that it might have would include: home Internet of Things data; personal activity data, e.g. location, exercise; home electricity usage; home device use, e.g. media viewing. There are several possible projects that could be done based on Databox (and others would also be welcome).

D1. Databox application

Identify a potential application area for Databox, recruit relevant stakeholders, and work with them to design, prototype and evaluate relevant Databox application(s). Prototypes may be of varying levels of fidelity (paper mock-ups through to complete functional Databox applications).

D2. Databox Application Audit

Databox is intended to support an open ecosystem of apps that use personal data. To support this ecosystem it is intended that self-selecting Databox users/developers will choose to scrutinise the actual behaviour of Databox apps in use, and their reports (e.g. rating) will allow other users to make informed choices about which apps to use. Design, prototype and evaluate one or more approaches (e.g. UIs, tool sets) for such auditing of Databox apps.

D3. Databox Security evaluation

Databox will implement a range of security policies and employ a range of security mechanisms, for example for authenticating users, managing accounts and devices, managing access to data. Security almost always involves complex trade-offs, for example with ease of use and compliance. Research and evaluate the initial approaches to security proposed for Databox; propose and evaluate refinements or alternatives of these approaches.

D4. Multi-user Databox

The first release of Databox will be single user; some initial UI sketches have been made for multi-user household use. Starting from these initial designs, design and evaluate approaches to managing security and permissions in a Databox with multiple users. Note that should consider issues such as sensitivities around highly individual data sources (e.g. one person's location or social media) on the one hand and shared/collective data sources (e.g. activity detected in a room, appliance and shared device use, shared cameras).

Mobile / adaptive music

Geotracks (http://dl.acm.org/citation.cfm?id=2967181) is an approach to adaptive music that tailors the listening experience for a specific journey. Initial experiments used an Android player, DaoPlayer (https://github.com/cgreenhalgh/daoplayer). In parallel, a "semantic player" application has also been developed (https://smartech.gatech.edu/handle/1853/54596).

G1. Geotrack Authoring

The initial geotracks were manually authored, largely by editing JSON configuration files. Design, prototype and evaluate a web-based system for non-experts to author geotracks and geolists (effectively location-based playlists). Variants of this project could emphasise adaptation of the playlist as a whole or adaptation of a single track.

General topics

In general am interested in:

Most of my current work is in mobile and ubiquitous computing (including mobile games and cultural experiences), and I have done a lot of work developing mobile applications on Android as well as web applications and collaborative systems.

Most of my projects combine (to varying extents) software development, interaction design and user studies.