FAQs

We are Johannes (tech) and Dorota (design).

Johannes has studied computer science (Dipl.-Inf. (FH)) and philosophy (M.A.). He has been working at a digital agency for 7 years, then freelance for the past 8. He lives in Berlin.

Dorota is a visual designer specialised in visual communication for knowledge transfer. She holds an M.A. in Graphic Arts and is completing a second M.A. in New Media. Since 2013, she has worked independently with clients across Europe, developing visual systems, interfaces, and communication strategies. She is based in Helsinki. Her work can be found at dorkastrong.com.

Our favorite clients are research groups, science institutions, and public organizations.

So far we have mainly worked with groups from agricultural and landscape research (list).

We make web-based apps and tools for science communication.

Our clients often want to reach out to practitioners in their domain of research, for example farmers and policymakers for an agri-economic topic.

Our apps and tools are designed to speak to those groups. We provide them with engaging ways to interact with, and benefit from, your research results.

For example, we put location-based data on fully customized, fluid map interfaces. We develop interactive decision-support features. We use illustrations, animations, interactive and multimedia elements.

We allow your users to gain insights that are customized to their situation. We enable them to provide information about their choices and preferences back to you.

This happens in intuitive, graspable ways, without your users having to learn anything upfront, and without them having to leave their device of choice.

We enable you to configure our apps and tools yourself: we either integrate them with your existing platform, or we provide a reliable CMS as the basis. This CMS can also be used to manage your project website.

The research projects we work with are typically public-funded. We are familiar with the processes, and can contribute to everything going smoothly on the organizational side.

We like opensource and opendata. We always provide you with all the code we write for you. We'll gladly publish it under an opensource licence if you want.

Finally, we are going to read your publications. It's one of the main reasons why we prefer working with research projects: it allows us to learn something interesting every time.

We adapt to the typical phases of a research project:

1. Planning: We help define the IT side of your project, and estimate costs and the timeline. We can also write a non-binding proposal to include in your funding application.
2. Kickoff and setup: Once the project has begun and your team is ready, we meet to discuss the current status, decide on requirements and main features of a first version, set dates for the milestones, and set up collaboration workflows that suit your team and institution.
3. Development: We build the app or tool in parallel with screen designs and layout sketches. Short meetings provide space for presentation, feedback and adjustments.
4. Testing and collaboration: We put working code on our server as soon as possible. We use simulated or preliminary data to provide an impression of the final product. Design progress is available for review and continuous feedback, typically using figma. All information, comments, and questions are collected in planning boards and a wiki.
5. Integration of research results: We prepare for when your research results become available. Integrating the real data into the app is a phase of close collaboration.
6. Handover: We transition the web app to your institution's IT infrastructure or to an external hosting service. You receive all the code, data and documentation. At this point, you will be able to make changes to your app's data and configuration without us. But if you want we keep backups, so we'll be able to support you in the future.
7. Perpetuation: We can set up a lean way to keep the project's online presence running after funding ends.

AI is currently a hot topic. We are reading and thinking about it, and have been for at least ten years. We are experts and very much up-to-date with the latest developments.

That being said, we have found that a useful way to think of AI, at least for now, is as a tool, nothing more. It may have big consequences (also in science), but it is not a paradigm shift.

Like other tools, it makes us better and faster, mostly in these two ways.

1. We use AI to write code when we already know what the outcome needs to be. In these cases, we specify the tasks, have AI do the implementation, and then verify and, if need be, correct the result. This saves time, because verification is often quicker than doing the implementation ourselves.
2. Other cases, when we are not as familiar with the task, also benefit, but in the inverse manner: we use AI to research the problem space and have it propose and explain aspects of the implementation, which we then do ourselves. Then we use AI again to review our solution and explain to us where it may be lacking.

Point being, we never blindly trust AI, but use it for autonomous implementation only in cases we can verify quickly. For the remaining cases, if we do use it to verify, we always have it explain to us how something can go wrong, to inform our decisions.

This is based on experience. We are constantly experimenting with different kinds of workflows, including full-on vibe coding as well as more restricted copy-paste editing without autonomous agents. The outlined approach seems to be a good combination of reliability and speed for the time being.

Our approach enables us to take responsibility for all that we build for you. This way, we gain some efficiency, while remaining able to give you estimates on changes, and propose practical features or solutions based on your requirements.

If you want, we can use local AI only, so that your code and data do not touch the cloud.