At the recently concluded ODECO Training Week 5, the entire ESR cohort participated in an ideathon, which is an organized one-day event where participants, often from diverse backgrounds, collaborate intensively to brainstorm, develop, and pitch innovative ideas or solutions within a specific theme or challenge. Unlike a hackathon, which typically focuses on building functional prototypes or code, an ideathon emphasizes the generation of creative concepts and strategic approaches.

As we were grouped into 4 different multidisciplinary teams, we needed to solve various problems related to open data sharing and accessibility. Our team was composed of Ramya, Georgios, Silvia, Louise and Gennaro and we were mentored by Els Breedstraet from the EU Publications Office. We tackled the following problem statement:

What technological solution can address the problem where non-governmental organizations are deterred from sharing open data due to: (i) ambiguities in licensing terms, (ii) a lack of consensus among stakeholders on licensing types, and (iii) being less trained in licensing compared to governmental actors who release more open data? “

It is known that dataset creators face the problem of plenty – there are many licenses to choose from, but there is no standardized workflow on how to make the choice. For example, the Open Knowledge Foundation maintains a list of more than 100 licenses that are compatible with the Open Definition. Els, our mentor, shared how the EU Data Portal identified more than 300 licenses that could be applied to make datasets open!

Having this knowledge, our team wanted to arrive at a solution that discharged two functions – one, to provide decision-support to dataset creators to choose the right license, and two, to serve a learning tool for dataset creators on how to navigate the different considerations involved in making this choice. At the same time, we wanted to be wary of a pure techno-solutionist approach to open data challenges, and wanted to infuse some criticality into our idea.

From all these design goals, our software-based solution emerged.

Architecture of TreeOfLicenses

At the back-end of TreeOfLicense, we envision a decision-support algorithm that can be visualized as an upside-down tree, with each variable branching into other variables, finally resulting in one or more licenses that fit the parameters of the dataset creator.

At the front-end of TreeOfLicense, users interact with a web-based questionnaire. Here, we took inspiration from other licensing tools such as the EU Data Portal’s Licensing Assistant. We noticed that these tools are quite static in their design – they either provide far too much information in a tabular form, or they exist as multi-step questionnaire where the respondent has no idea about their progress in the questionnaire or about how the questionnaire itself is structured.

Design requirements

Because the first design requirement was to design a more interactive, maneuverable interface, we first decided to design a simple questionnaire to be answered by a dataset creator (or even re-user). This questionnaire would ask the respondent to reflect on the source/provenance of each individual data element in their dataset, their objectives with creating and sharing the dataset, and their proprietary/other interests in the dataset. Each response would be mapped on to the back-end decision-support algorithm, to generate a list of licenses that meet the respondent’s parameters.

The second design requirement was to allowusers to be able to “see” the internal workings of our decision-support software. This enables our software to serve as a learning tool for respondents which allows for more transparency of our software, and moves away from a techno-solutionist approach. We this decided to create a user interface, where the respondent can always see each section of the questionnaire as well as their progress in the questionnaire as well as be able to toggle between questions and alter answers.

The final design requirement was that the license decision tree allows users to interact with the tool in a dynamic way. We addressed this requirement by designing the license decision tree to regenerate after each response and modification of a response. Furthermore, instead of being visible to the respondent only at the end of the questionnaire, this dynamic license tree is always visible to the user as they go through the questionnaire.

How did we do at the ODECO Ideathon?

Our idea was well-received during the Ideathon by both supervisors and external assessors who were part of the ODECO Training Week. We won second place!

We believe that solutions like this are necessary to provide more clarity to dataset creators, as well as re-users. And if we receive some seed funding from one the readers of this blog, we could be well on our way with TreeOfLicenses to address a significant challenge in the open data ecosystem!

Acknowledgements: TreeOfLicenses was conceptualized by the three of us together with Louise van der Peet and Gennaro Angiello, who were participants at the 11^th OpenGov Summer School conducted in parallel with the ODECO training week. We would also like to express our gratitude to our mentor Els Breedstraet for her valuable guidance. 

Authors

Ramya Chandrasekhar, CNRS

Silvia Cazacu, KU Leuven

Giorgios Papageorgiou, Farosnet