The real story behind our food is complex and invisible.

For many of us, our everyday experience is a repeated journey of planning, purchasing, cooking, and eating—or of arriving at a restaurant, ordering, waiting, and eating. Thanks to food systems that operate behind the scenes, many of us don’t worry too much about where our food comes from. The unintended consequence of this hidden complexity is when major disruptions occur, like climate change-induced shocks or war, we don’t know how they will affect our access to food, not to mention the rippling impacts to those far removed from the actual events. On a larger scale, how our food systems function makes it difficult for international aid groups to plan resources and funding adequately. While there are forecasting tools for catastrophe scenarios, few tools are trying to monitor the whole food system.

Shedding Light on the Global Food Landscape

This challenge inspired us to collaborate with Earth Genome and Zia Mehrabi’s Better Planet Lab at CU Boulder to help visualize the global food system. We built on an earlier iteration, the US Food Twin, aiming to now create a global scale Food Twin. We focused on creative and impactful design that would elevate collective awareness of the worldwide production, consumption, and trade of food.

The sheer scale of global food systems presents a formidable challenge in data modeling and representation. While we were eager to tackle this with our partners, we first needed to address a fundamental question: how do we conceptualize and visualize a system of this complexity in a way that's both compelling and true to the data? This challenge was at the heart of our project, driving us to nail down an approach that could handle massive datasets while fulfilling our ambitious vision. Only then could we move forward with crunching the numbers and creating an accurate model.

Remote design sprints have been a hot topic since the pandemic forced many teams to work virtually. While there is lots of advice and suggestions on the basics of running remote sprints, our experience with the Global Food Twin offers a unique perspective of this process for complex, data-driven projects.

A Global View of Food Systems using Collaborative Design

We decided that a design sprint with Earth Genome and the Better Planet Lab would help to define a shared product vision for this global food twin application. A design sprint is a time-constrained, collaborative approach to solving a well-defined challenge, resulting in a tested (or at least testable) and actionable solution.

Over the course of several weeks, we undertook a series of workshops inspired by a design sprint methodology to work through the most ambiguous part of the problem together.

However, the globally distributed nature of our teams presented another novel challenge for us to solve – how should the typical in-person, 5-day design sprint be adapted to accommodate our team's time constraints? We made several adaptations:

Longer Duration

Instead of consecutive days, the sprint spanned several weeks. Indeed, a “sprint” implies a short burst, and this was still an expedited phase in the context of a many month project timeline. The gaps between sessions also allowed for reflection and deeper thinking between activities.

Setting up a live collaborative space

One benefit to remote sprints is that additional team members who may not have been able to participate in person can still reference and participate in digital spaces as the project progresses. We organized these spaces in a way that allowed everyone to use them asynchronously. Clear instructions and pre-populated content on digital boards and documents are keys to success.

Record Everything

With any remote workshop, there will inevitably be time zones and scheduling conflicts that don't allow all participants to fully join all sessions. Identify the 1-3 key people who must be present for any given session and be flexible so that others can watch the recording of the sessions they may have missed.

A remote design sprint allowed us to pivot the agenda as needed, and we were able to close out the sprint with excitement and momentum.

A Step-by-Step Journey Through the Design Sprint

Phase 1: Scoping

This phase focused on defining and refining the project scope. We began by looking at the application's potential users and use cases, encouraging everyone to start thinking from a user-centered perspective.

Discussions about people and real-world problems can often be overlooked in light of excitement around novel data, technology, and solutions. Prior to the meeting, we asked team members to share thoughts on who they imagine this application should serve.

There were about ten different use cases discussed, of which three main user groups were identified that centered around preparing and responding to world events that disrupt the global food system:

Phase 2: Understanding the Current State of the Data

The next session was a chance to align on what data we had and what data would need to be generated.

For this project, data was related to global food producers, consumers, and the flow of food between them (e.g., supply chain). Additionally, we needed to be able to model disruption scenarios due to factors like conflicts, trade agreements, and export restrictions. We spent the most time discussing the data around disruptions as it was the most open-ended with many possibilities.

After the data discussion, we brainstormed potential user questions that the data could answer through an interface. Some of the top user questions the team wanted to visualize with the data were:

• What does the global food system look like?
• What risks do we not know about that this new view can tell us about?
• In the early days of an event, where monitoring is tricky, who are the people likely affected by a global or local event?

Phase 3: Sketch

Instead of starting from scratch, we wanted to build on what worked well in our previous work on the US Food Twin and what could be improved when moving to a global scale. The team individually brainstormed the positives, limitations, and recommendations for the Global Food Twin regarding the UI/UX design and application architecture.

We also reviewed existing applications, visualization techniques, and websites that could be remixed and improved for the Global Food Twin. These were gathered in advance of the session to save time doing online searches while meeting together.

Example applications or visualizations for inspiration included:

• Surfacing stressors (e.g. Port Watch and FEWS NET)
• Sankey diagrams
• Visualizing transport (e.g. Resource Trade and Crop Monitor)

In our collaborative workshop we invited everyone to contribute to the design directly by sketching out their ideas. Everyone spent some quiet time sketching their interface ideas either on paper or digital drawing tools. Given the complexity of the visualization, individuals focus on a single feature or general overall design idea. Doing this digitally removed the barrier of having to feel like an artist – people could use drawing tools they were comfortable with, so it wasn’t about how well you could sketch.

Here are some of the highlights:

Initial Wireframes

After the two sprints and multiple sessions, our designer, Fausto, took the lead in producing initial wireframes. The initial wireframes organized the application into food system flows, disruptions, and insights. Countries were used as the key interactive element because they are administrative areas that people can understand. By selecting a country, you would be able to see the flow of food moving from the country's main hubs to other countries.

The main feedback in the first round was to consider a more fine-grained way to interact with the data than just countries. Even though countries are recognizable areas, there was a lot of data within a country that could offer insight into areas that produced comparatively more or less food.

Iteration and Final Vision

At this stage, we decided to reimagine the user experience completely. Our goal was to create an insightful and comprehensive app, delivering a full data experience while maintaining the simplicity of interaction found in the US Food Twin.

We envisioned a clean interface without any menus or buttons centered around a slowly rotating globe. This globe would highlight three main categories of interaction: the production areas, the transportation nodes (such as ports, airports, and shipping hubs), and the shipping routes (by land, sea, and air). Users could simply click on an item on the map to begin exploring the data.

We proposed a scrolly-telling approach to keep a simple interface while also telling a story. As users scroll, the data on the map and panel dynamically updates to show different widgets of the food system's journey—from production areas and shipping nodes to shipping routes and their impact on people. These widgets are the building blocks for our interface, allowing us to abstract and repurpose these elements across the three different starting points. This strategy allows us to provide a cohesive user experience and makes the development process easier.

Turning Plans into Action

The Global Food Twin sprint was a great example of how design sprints can help our partners, users and other stakeholders build the right solution for the right problems, increasing value and potentially decreasing long-term project costs.

The concept of remote design sprints continues to evolve as remote work has become status quo for many organizations (like us!). Our experience with the Global Food Twin project demonstrates that these methodologies can be successfully adapted for long-term, complex projects involving multiple stakeholders and large datasets. Design sprints can be adapted in many ways to meet the constraints or unique needs of the project team. In this case, a fully virtual sprint spread out across weeks allowed us to accommodate a distributed team. This gave us the added benefit of time in between sessions to let ideas simmer, and give designers space to create wireframes for early feedback. We can confirm: Design Sprints lead to better products, and often cost less in the end.

While many other design teams have paved the way for remote design sprints during the pandemic, our approach shows how these methods can be scaled and customized for ambitious, global projects. By combining the intensity of sprint methodology with the flexibility of extended timelines, we've created a hybrid approach that maintains creative momentum while allowing for the deep thinking required in data-heavy projects.

We are excited to continue this collaboration and bring more life to the application in the months to come. You can follow Earth Genome, Better Planet Lab and us to stay tuned when the application launches. If you think your project could benefit from our design sprint process, we'd love to talk with you.