Media Monitoring

The Media Monitoring use case focuses on the observation of web feeds with media content. The results of the analysis can be visualized in dashboards and used to trigger alerts.

In the Media Monitoring Use Case, there are two use case applications:

    Advanced Content Analysis

    This use case application covers advanced analysis of topics of journalistic content and summarization, which require more research. This may include a wide range of topics and themes.

    Example: Broadcast Monitoring & Analysis

    The goal is to enhance internal and external monitoring and analysis activities based on reallife requests from within the broadcast (in particular DW) community, using advanced SELMA research results.

    During the project, DW will apply the SELMA platform to monitor broadcast material, from DW internal sources as well as from other news providers. This covers ingestion and processing of massive data streams, with video, audio and text, including feeds in all 30+ languages from DW. Advanced content analysis supports a more efficient editorial workflow and decision-making process through better information on trending topics and breaking news. It is a major step forward towards breaking the language barrier, as awareness of and access to content from different languages brings the different language departments within DW (and by extension different communities) closer together. In addition, ad hoc, specific analysis requests will be done based on topics of interest reported by DW users, e.g. elections or high-profile events, statements or actions by specific politicians. NLP technology enabling high-quality, reliable filtering and clustering, resulting in cross-language summarized reports allows our editorial staff to work at a next level in a fully customized way.

    Example: Diversity

    DW proposes to analyze journalistic content with respect to diversity, a major topic of interest, high on the agenda in the media as well as in the society. It is a complex analysis, thus well-suited for advanced analysis applications.

    Media needs and wants to be representative in their journalistic output.

    “An editorial team loses credibility if it doesn’t represent diversity in its many forms: gender, ethnicity, sexual orientation, disability and different world views. Credibility is the most important currency in journalism” (Peter Limbourg, Director General Deutsche Welle).

    We intend to improve diversity and representation in media by counting numbers and appearances with the help of Text Mining and NLP. This will be done in 30+ languages (using machine translation), covering the range of DW languages.

    We will measure both text and audiovisual content, published on and beyond.

    Automatic reporting will make the analysis output available in the form of summarizations and graphical visualizations.

    Press Agency Analysis

    This use case application, led by Priberam, targets a particular user set of Media Monitoring platform users, Press Agencies (PRs). Typically, these users serve a set of clients, operating in different areas of business.

    To specify the monitoring activities, the user sets their preferences, which may include which topic areas should be monitored (specific entities/organizations, different topics – sports, politics, economy, etc.), which sources can be shown in their original languages, and which should be displayed in English (default) or another chosen language. All these settings are saved in a so-called “View” – in other words: how you look at things – which can be reviewed and adapted at any time in a dedicated area.

    Based on these settings, the system provides search results and other visualizations which the user can view either as a larger cluster or as individual items. Both clusters and items can be edited or adapted while working with them: users can mark items as favorites, save or curat them, provide feedback to the system, mark them as read or unread, get translations or put them through the fact checking process. These adaptations and configurations are used to train the system to improve following activities of the monitoring process.

    Based on all these activities, the user can produce various outputs, like reports, dashboards, etc., and translate them into any desired language supported by the system. These outputs can either be shared internally, e.g. with colleagues or the head of department who had asked the analyst or journalist to do the monitoring, or they can be exported to PDF to share with clients.

    Bring me to News Production Use Case.