El periódico

Personalised news services are already able to help people find news that is relevant to them, to recommend the right news to the right users when they want it, and to help users keep abreast of news by aggregation over multiple sources. According to Billsus and Pazzani (2007) the adaptivity in news access is achieved through several methods:

• News content personalisation (refers to content): Push filtered articles pre- dicted to match the user’s interests.

• Adaptive news browsing (refers to the user interface and interaction): Change the order of categories of articles.

• Contextual news access (refers to the user interface and interaction): Offer users access to additional information related to the news they are reading. • News aggregation (refers to content): Automatically identify main news top-

ics emerging from multiple sources.

News consumption on the go has now become the priority for the leading news organisations (Kelion, 2015). The diffusion of smartphones has changed the way of news production and consumption. App marketplaces are already bursting with prominent apps for accessing news spanning the globe, delivering completely tai- lored news from multiple sources and offering the chance to share news content in social networks. Even more news portals and websites are optimised through re- sponsive designs to support the mobile experience. A closer look at how smartphone apps and websites achieve personalisation, however, suggests that the majority of them, specifically news apps, allow users to manually create a personalised experi- ence. For example, users can explicitly select topics of interest or specify system

parameters on how they want the visual presentation of a story or how the stories are organised.

Surveying news apps from Apple’s and Google’s marketplaces that provide personalised experience will shed some light on adaptivity in this very specific do- main. It will provide a better understanding of how personalisation is achieved and, of course, it will identify possible gaps that would inform the design of more personal user interfaces. The review reported in this section is based on online tech blogs including the DigitalTrends, the Wired, the BusinessInsider and the Sim- plyZesty. In addition to reviewing commercial news apps, the section reports re- search works and prototypes that have been proposed.

Leading news organisations such as BBC and CNN have already realised the need for personalization in their own news apps. For example, BBC news app provides a more personal news reading experience through customisations of the interface and other system parameters related to the content. Example features of the revamped app include the most read stories, an option to add a list of news stories a user follows, presentation settings of displaying and categorising the stories such as a compact layout or carousels and many others.

News aggregators are another breed of news apps in which the service mainly focuses on the aggregation and classification of news content from multiple sources. As more news sources are emerging with a tremendous number of stories spanning all over the world, news aggregators help users to identify news topics of interest easily and to access news topics from different news providers. Flipboard 3, for example, uses the metaphor of a ‘personal magazine’ by making the entire read- ing process stylish. It gives the sense of flipping a magazine page while navigating through news. Users curate and share their own mini-magazines with the app, draw- ing in stories on their preferred topics. Zite 4 is an intelligent magazine-like news app that recommends stories based on users interests and reading habits. The app learns users preferences through a thumbs-up or thumbs down button on each story.

3_{Flipboard: https://flipboard.com/}

2.3. News Consumption and Adaptation in News 29 Inside.com Breaking News5allows users to select news topics to follow and then provides 300-character summaries of relevant stories along with links to the orig- inal sources. Newsbeat 6 is another aggregator but one that creates ‘personalised radio news bulletins’. Users select their preferred text news sources from which stories are pulled each day. Summaries are created and in turn news podcasts are generated using text-to-voice technology. Feedly 7 aggregates news items, longer articles, blog posts, and quick videos into a single spot in an elegant way. Fur- ther, instead of providing a massive list of articles it breaks the content feed up into manageable chunks. News360 8 differentiates from other aggregators by in- corporating two ‘swipeable’ screens, in which the top part shows the most popular stories while the bottom displays the current article you are reading. Social net- working platforms such as Facebook and Twitter are becoming distribution chan- nels for news stories. Recently there appears to be a huge interest in such services with more people getting their news stories and updates from social networks; as numbers indicate (Reuters Institute, 2015). This kind of service, therefore, could be used to develop apps that pull or leverage knowledge from users’ social net- work activities. For example, Pulse, developed by LinkedIn, delivers personalised news from a user’s professional network. Phelan et al. (2009) proposed a system coined Buzzer, which recommends and ranks news articles by analysing real-time Twitter data. Further, LumiNews (Kazai et al., 2016) is another mobile app proto- type that provides personalised content recommendations by leveraging Facebook and/or Twitter feed combined with a users current location to automatically infer that user’s news interests.

Apart from news apps, web portals such as Google News and Reddit aggregate news sources and/or recommend news articles to assist desktop end-users to find and read news more efficiently. These systems gather information about their users either explicitly, i.e. users give rates to articles, in the case of Reddit, or implicitly by observing user behaviour, i.e. track users’ activity, reading preferences, etc., in

5_{Inside.com Breaking News: https://inside.com/} 6_{Newsbeat: http://www.bbc.co.uk/news/newsbeat} 7_{Feedly: https://feedly.com/}

the case of Google News (Liu et al., 2010).

Although a large proportion of news apps on marketplaces (Google and Apple) adopts the adaptable way of providing personalisation, the alternative is the use of adaptive principles. Adaptive systems attempt to overcome some of the limitations of adaptable systems by automatically personalising the user interface, without any manual user intervention. They mainly leverage prior knowledge about the user to infer their goals and needs to automatically alter the system’s behaviour. However, despite these potential benefits of adaptive systems, news apps tend to adopt the adaptable principles by allowing users to manually customise the content or the in- terface. However, the evolution and adoption of technology (Pew Research Centre, 2017b) have transformed they ways users interact with user interfaces. Adaptive principles could possibly work better for today’s smartphone user interfaces. To- day’s smartphones have much more advanced capabilities such as 4G connectivity (i.e., connected anywhere, that might increase user’s experience when interacting with UIs), high-resolution screens, sophisticated interactions with the user interface (swipe, flick, scroll) and others.

Research-oriented prototypes for news personalisation have also been pro- posed over the last decade, mainly for web or desktop applications. A few prop- erties these systems share are the aggregation of multiple news sources into one application, filtering and ranking of news articles based on users’ feedback and fi- nally recommendation of news articles based on users’ interests and preferences.

Billsus and Pazzani (1999) developed NewsDude, a news recommendation system, aimed to recommend news articles for desktop users. News Dude relies on feedback from the user to automatically adapt to the user’s preferences and interests. It uses a multi-strategy approach to model a user’s interests in which a nearest-neighbour algorithm is used to model short-term interests and a Na¨ıve Bayes classifier for long-term interests. Carreira et al. (2004) used interaction logs of mobile users of news services to implicitly capture user profiles as the basis for recommending articles of interest. Their prototype news application developed for PalmOS-based PDAs logged aspects of users’ reading behaviour such as reading

2.4. User Modelling 31