#	Pagina
attuale pagina	/open-h2020/projects/211241/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - MARCONI (Multimedia and Augmented Radio Creation: Online, iNteractive, Individual)

Teaser

Summary

Nowadays, radio shows surpass linear broadcast feeds by increasingly revolving around user engagement, which encompasses audience interaction and experience personalisation. Unfortunately, the tools to enable radio-makers to capitalize on these new kinds of media and interaction opportunities are premature at best, and are in addition ill aligned with contemporary radio production workflows.
MARCONI will enable fully interactive and personalised radio formats, integrating broadcast radio with digital and social media, resulting in converged radio experiences. To realize this goal, MARCONI will pursue two concrete objectives. First, consumers will be able to interact with â€œliveâ€ radio through their preferred communication channel in ways that are richer than is feasible today. As such, consumers will feel more connected to the radio brand - even when not listening linearly - and better served in an individual way. Second, radio-makers will be given an integrated view on audience interactions and will be supported by interaction automation services. The net outcome will be enhanced audience engagement.

Work performed

From project vision to use cases, requirements and prototypes

The project departed by analysing a first set of scenarios captured in co-design workshops. This analysis served as a basis for exploratory observations at the radio stations and a second ideation round in the form of co-design workshops. Then, an analysis exercise that considered the work of all ideation rounds resulted in the translation of the insights in user scenarios, technical requirements, privacy assessment and system architecture.
During prototyping the work has been focused primarily on designing and evaluating exploratory mock-ups, and on initiating the functional prototyping activities. Piloting constitutes the next logical step in MARCONIâ€™s user-centered design approach. MARCONI pilots are intended to verify the usability of the MARCONI concept as a whole under (semi-)controlled usage scenarios over prolonged periods of time. All three broadcasters have been involved in these activities, for example by including chatbot services into existing apps and websites, and connecting MARCONI APIs (e.g. for messaging and search) with the editorial tools in place. MARCONI harnesses a continuous validation and evaluation process, which applies to mock-ups, functional prototypes and pilots.

Development of the minimum viable product (MVP)

For the MARCONI project, we chose to work with an agile development method. While there are several known agile management methods available, they mostly require a small team to work closely together. To adapt to the needs of MARCONI, we used several tools and techniques from known methods such as XP, SCRUM and Kanban, but applied them in a different way.
The objective of the initial development was to arrive at a minimum viable product (MVP). It is a minimal working version of the MARCONI platform and services. The MVP has been the first focus area in the project development, as it corresponds with a milestone for the start of piloting activities. By creating demos, consortium members are able to get a clear view on what the end result might be. In addition, making things tangible sparks-off discussions leading to further ideas. A scrum sprint is a regular, repeatable work cycle in scrum methodology during which work is completed and made ready for review. The sprints in MARCONI are between two and four weeks in length.
The systems are built as platform services separated into microservices. Different systems are controlled or implemented by different MARCONI members. With the separation of services we can create different solutions adapted to the situation in every broadcaster. The ultimate aim is to integrate with production systems at the broadcasting partners. Such a flexible architecture is a necessity for achieving this aim.
The sprints integrated the intelligent services developed so far, which include chatbot services, search in social media content and image analysis. Furthermore, the services, connect to existing applications or websites at the different broadcasters as well as RadioManager as the core platform.

Privacy aspects

An in-depth analysis of the architecture, data types and proposed services of MARCON from a legal point of view has been performed. Each dataset is classified, if it contains personal data according to the GDPR and under which conditions such data can be processed lawfully. In MARCONI, this shall either be consent, a necessity for the performance of a contract (smartphone app) or the purpose of legitimate interests pursued by the controller. The data being collected in the planned pilots have been described systematically.
As part of the MVP development, a privacy platform called PriVaults has been developed. It serves as the central hub for accessing data, and allows formalising data processing purposes and permissions, so that they can be checked against registered consent from users before components of the system are granted access. PriVaults has been the result of comprehensive dis

Final results

A key innovation of MARCONI is the PriVaults platform, which provides a state of the art solution for controlling user data across the entire ecosystem, including services and apps from third party developers that are integrated with the system.
Further technical innovations have been made in the development of intelligent services. Conversation services such as chatbots, which currently rely on static knowledge bases have been connected with dynamic information sources, such as programme and music information from RadioManager. The face recognition approach has been modified to enable fast online training starting with only few samples.
The expected results until the end of the project is a platform for radiomakers, which can be connected with a broad range of intelligent analysis, as well as data and communication services. The services are provided as microservices, which enables their integration with existing infrastructure at broadcasters. The development of the platforms and services strictly follows a privacy by design approach, and thus gives consumers full control of their data while maximising the benefits of data analysis for customisation and personalisation of content. These services will enable better interaction and personalisation for media consumers, and will increase the use of state-of-the-art IT services in media production.