On the pulse of science with KATI

From the idea to the program – a look back at almost seven years of KATI

As part of its technology foresight activities, Fraunhofer INT has established a kind of science observatory. This means that all fields of science and technology are continuously monitored in order to become aware of new technological developments as well as breakthroughs in already identified topics. The central source of information for this process are scientific publications and patents. When the topic of cognitive computing was hyped a few years ago, Fraunhofer INT came up with the idea of utilizing this approach for technology foresight. This was the birth of the KATI-system.

Today, several years and many hours of development, programming and testing later, the initially rather vague idea of KATI has become an important tool for the daily work of scientists at Fraunhofer INT. And KATI is not only used there – since last year, the system has been available to all employees in the Fraunhofer-Gesellschaft. In addition, with the support of funding from the Federal Office of Bundeswehr Equipment, Information Technology and In-Service Support (BAAINBw), a new field of research has emerged. This new field, focused on data driven foresight, is the focus of the KATI Lab and addresses the central question of how foresight processes can be made better, more effective and more insightful through the use of data. In a certain sense, KATI is a tool for data-driven foresight that focuses on publications and patents as data sources.

 

Data – lots and lots of data

When KATI was presented to the wider public for the first time in August 2017, the database contained the metadata of almost 60 million publications. In the course of 2024, the mark of 80 million publications will be exceeded. An important design decision was to make this data accessible using a graph database. The data model, which was developed by the KATI team at a very early stage of the project, has since then proved very successful. On the one hand, it can be adapted for other data sources. These include, for example, publication data from Dimensions or patents. On the other hand, this approach simplifies a large number of analyses and facilitates explorative and interactive data analyses.

As part of the development of the KATI-system, the team is focusing on close integration with specific questions and use cases of the users, as they arise from the daily work in the various projects. The focus is therefore less on exploring additional possibilities and more on identifying what is needed to support a foresight process as efficiently as possible. This has an impact on several levels.

 

The development process 

First of all, this results in specific analyses that are implemented in the system. One such exemplary use case is the question of which publications you should read in order to familiarize yourself with a topic as quickly and comprehensively as possible. The reference citation plot was implemented to identify such so-called key publications and has proven to be very useful. Other use cases relate to topics such as actor analyses or the identification of potential new fields of application.

This use case-specific approach is further supported by the flexible user interface. This makes it possible to compile individual perspectives on a specific topic. For example, if the research focus of a particular institution is to be analyzed, the corresponding bar chart can be used as a filter to take a closer look at the keywords for the publications of an institution and thus generate an insight into its research focus.

Finally, the needs of the users also play a central role in the design of the graphical user interface. In addition to an appealing design, the KATI team's development work focuses on simple and intuitive usability. The aim is to ensure that the KATI-system provides its users with as much added value as possible. A current example of this is the implementation of a query editor, which offers the visual option of using the KATI-system's comprehensive search options via drag & drop. In this way, users are able to design complex search queries simply and intuitively.

 

Future prospects

The KATI-system will celebrate its seventh birthday in 2024. However, development is far from complete and is constantly being driven forward. In addition to the continuous expansion of the methodogical portfolio, the focus here is on tapping into additional data sources. And with large language models, there is already the next hype in the field of AI research, the use of which for technology foresight represents an exciting scientific issue that is also important for the development of the KATI-system.