Work Packages

We’ve defined the following 7 work packages which reflect the orientation of our research project in more detail.

• Problem Analysis & Workflows
• Visualization and Interaction Components
• Automated Analysis Components
• Application Scenarios & Integration
• User Evaluation
• Dissemination
• Project Management

1. Problem Analysis & Workflows

We pursue a thorough understanding of the contemporary challenges in DDJ in terms of user, data and task characteristics, as well as current practices, problems and stumbling blocks. Investigations to address those issues will include close collaboration with domain experts in order to discuss and define currently unsolved problems. Our problem analysis phase will focus on:

• evaluation of currently used methods and applications in data journalism
• data availability and processability
• gender gap issues
• user needs, tasks, and goals (our users are data journalists)

According to our findings we will derive implications for design, technique and tool development as well as guidelines and best practices about how to integrate DDJ into newsrooms and how to adapt journalistic education.
Our methods to achieve these goals will include document and domain literature reviews, semi-structured interviews, contextual inquiry, requirement analyses, surveys, newsrooms observations and focus groups with journalists.

2. Visualization and Interaction Components

This work package aims at the design and development of novel visualization and interaction components. Resulting software components will offer means for Visual Analytics of time-oriented textual data (e.g. unstructured text like pdf files) and dynamic network data (data reflecting relationships between any kind of entities). We will put specific emphasis on usability, understandability, and learnability of our tools and techniques.

After assessing existing techniques and scouting available software libraries and toolkits we will start the development of interactive visualization tools aimed at supporting data journalists in finding the hidden stories within their data. Running prototypical software will be provided as proof-of-concept, to be evaluated comprehensively by end users (see work package 5). Scientific and practical findings of this work package will be presented continuously at national and international conferences/workshops as well as published in selected journals (work package 6).

3. Automated Analysis Components

Work package 2 and work package 3 will work closely together to achieve an integrated Visual Analytics approach. The visualization and interaction components implemented within the scope of work package 2 need an underlying layer focusing on lower-level information processing issues. In this work package we will assess the state of the art of methods and measures from the fields of natural language processing (NLP), graph theory, statistics and hybrid coupling.

The elaboration of those fields, complementing interactive data visualization, will be used to build upon existing approaches and estimate the trends of future research. We will extend current methods for text- and network analysis and develop new ones. Those new methods will be implemented as proof-of-concept prototypes. The development of our prototypes will be based on Rapid Prototyping and Test-driven Development (TDD).

4. Application Scenarios & Integration

To analyze DDJ-specific workflows we will define two real-world scenarios based on two publicly available data sets:

• (1) analysis of Austrian parliamentary speeches from almost 20 years (large collection of unstructured and semi-structured stenographical protocols as pdf files)
• (2) investigation of direct and indirect governmental advertisement in media (structured dataset reflecting relationships between Austrian politics and economy)

The software components developed in work packages 3 and 4 will be integrated, validated and refined towards these scenarios.

The development streams of software tools will be consequently monitored by our DDJ-experienced commercial partners to minimize the commercial risk when entering the market with a product in the future. To adequately address our target group’s needs we follow a user-centered design approach based on problem analysis and design requirements (see work packages 1 and 5).

5. User Evaluation

To continuously and rigorously evaluate our proposed solutions, we will focus on two different streams:

• formative evaluations will be conducted with the goal to further improve our proposed solutions,
• summative evaluations aim at showing the effectiveness of our solutions compared to other state of the art techniques in the field.

For formative evaluation we will use a broad range of usability engineering methods which are standard practice in the human computer interaction community. These include but are not limited to:

• cognitive walkthroughs,
• heuristic evaluations,
• and think-aloud studies.

We will also engage in different summative evaluation methods, mostly stemming form the research fields of visualization. For evaluation of our technical contributions (work packages 2 and 3) we will conduct case studies with real datasets, as well as user studies in the lab. For the application scenarios (work package 4) we will focus mostly on field methods, studying our solutions with real users and real data “in the wild”.

6. Dissemination

The goal of this work package is to broadly disseminate the work conducted and results achieved to different stakeholders such as scholars, journalists and interested citizens. Part of our dissemination strategy are tight communication and collaboration with researchers, data journalists, media agencies and all other people interested in our activities. Furthermore this work package covers our project website, our project blog and our newsletter. As part of our social media strategy we are present on Twitter as well as Facebook. The source code of our software prototypes will be publicly available on our GitHub account.

All researchers involved in our project will present papers at scientific conferences and publish articles in scientific journals. Priority will be given to open-access publications and world-leading journals and conferences. For Visual Analytics these are in particular IEEE VAST, IEEE InfoVis, TVCG and ACM SIGCHI. For communication science these are New Media & Society; International Journal of Press/Politics; European Journal of Communication; Science Communication; Journalism Studies; Information, Communication & Society.

Members of all consortium partners will also present the project to relevant application communities. In particular, we plan to hold an empowerment workshop series in which we will teach journalists basic DDJ skills. As a measure to tackle the immense gender gap in DDJ, two of the four trainings of this series will be held by female instructors and will be only open to women. Furthermore, we will give presentations at Austrian media organizations and also at international journalism events such as Deutsche Medientage München, re:publica Berlin or the International Journalism Festival Perugia.

7. Project Management

The purpose of this work package is the efficient planning and coordination of the project to meet the project goals, ensure overall project quality and stick to the time schedule. The scientific and organizational tasks of the partners will be coordinated, planned, and monitored to accomplish a high quality of interpersonal collaboration. The most important phases of this work package are initialization, planning, controlling, risk management, evaluation, the adaption of our work plan, and the successful completion of the whole project. The parallel development strands of work packages 2-5 will be synchronized and coordinated particularly at given milestones as well as in our quarterly consortium meetings.