IRFS Weeknotes #113
While the weather continues to be unexciting, our team has been very focused on taking forward several strands of work and hoping for a better summer :-) Let's start with one of the main projects the team is "sweating" on this summer.
This week we have officially started our next phase of research work for the FI-Content project but let's first recap what has been achieved in the past month as part of our research on privacy and user data from TV consumption.
We have explored users' awareness of activity data stored by the service. Did they know this was possible on TV? Are users ready to allow services to track their attention data in order to receive recommendations or maybe track where they are into a series? Do they trust public broadcasters to hold their data? We approached it by putting user data transparency and control in context - what are the benefits?
We tackled the research questions around transparency by developing an online prototype of a data dashboard for a closed trial, fed with information gathered from the diary study we conducted with the participants over a period of two weeks to record their activity on TV. Each participant's dashboard data was presented back alongside a series of paper prototypes showing some of the benefits of data gathering (i.e. progress in a programme and recommendations) in the context of a BBC product. Some screenshots were posted in an earlier blog post.
From the evaluation emerged the most popular format of user data history was the thumbnail view because it is easier to identify what programmes were watched and also because pictures were a memory aid. The majority of the participants seemed unconcerned if their TV attention data was tracked however there was a considerable percentage of users who would want to consent to this and have an awareness that this was happening, hence the need for simple, clear information upfront. Some participants to our study were very surprised to learn set up boxes they already subscribed to have access to their viewing history but were not aware of it as such information is possibly buried under pages of T&C on their contract. Users also said they would trust their data being held by the public broadcaster and it is such trust we need to ensure it is kept high by assuring users data is held securely.
Our conclusions on the user data transparency study were that participants were largely positive about the potential improvements to their TV experience through the capture of their attention data and were mostly comfortable about their TV data being collected as long as they trust what the service provider will do with it and have a degree of control over it by offering the option of limiting that use or stop it if they wish to.
In order to investigate to what extent users would like to have control over their data, we have developed a Chrome extension prototype of a viewing history dashboard on the iPlayer.
The extension allows users to view, delete, stop tracking (via private browsing mode) and export their attention data.
Participants to this study have been using the extension for a period of 30 days ending on Monday. Their interaction with any of those features has been logged and the analytics will help us evaluate what features users would use the most, in what context and which ones aren't used and why with the help of a end of study questionnaire we have prepared. The report that will follow from this will complete our study around privacy as part of the FI-Content.
As mentioned before, this week we have started another strand of work exploring ways of authenticating a connected TV to an online account with the objective of developing a protocol for authenticating a connected TV to an online account but also to continue working on the egBox (streaming live video in HTML5 research platform) in order to support authentication. Dan and Sean have started by updating and cleaning up the egBox ready for implementing a RadioTAG client next week. While Theo and Joanne have spent most of the time locked in a room mapping out interaction models for pairing devices and the user's profile.
In parallel to those two strands of work, I have started to plan how to tackle the numerous project's deliverables due between July and September. Not great fun writing EU reports during summer. Chris Needham has instead been at the FI-Content plenary meeting at DFKI in Saarbrucken in Germany. At the meeting the main topics for discussion were how the FI-WARE project is progressing, in terms of delivering software components and services that we can build upon, and how we would organise user testing in the next phase of the project. We also planned which of the technologies we've developed would be shown during our workshop session at the next NEM Summit.
Our team's other big workflow this summer is on the ABC-IP project with Andrew N. this week working on public prototype, which is taking shape. We have now basic search results, a home page with a featured programme and a basic audio player. The Tagging Experiment that we were running with Global Minds is now finished and the response to the survey has been fantastic and now it's time to analyse the data. Pete had also joined the ABC-IP team this week beginning wireframing the new UI. Chris L. was also working on the public prototype of the World Service archive building a small ruby library to fetch current news articles from Wikipedia and started relating episodes in the archive to the news events. Duncan has been creating a Solr index of World Service data that we can use to power the initial phase of our web interface.
Yves has been mostly working on some features of the public World Service prototype. He has also been doing speaker segmentation and identification on the From Our Own Correspondent archive. He has also been following up on contacts made over the last couple of weeks at the Trade Fair in Salford and at the Europeana meeting in Leuven. His slides are now available. He has also started extracting topics from a large archive of subtitles as well, which is giving fairly interesting results.
Olivier instead was in Brussels this week for a workshop on Open Data (policy modelling, transparency and data journalism were among the main themes) and presented alongside Andrew Leimdorfer of the News Specials team about our work on visualisation, sense making and linked data. A nice summary of the event can be found here.
In Manchester Michael Sparks as part of his IRFS time has been primarily interviewing candidates with Vicky for the Internet of Things Salford University fellowship.
Thanks to Olivier, George and Akua who were away for most of the week we got lots of sweets on our cake table on Thursday. That's what I call happy ending :-)