How to Visualize Complex Systems – Conceptualizing Global Food Insecurity

via Jalal Awan, Patricia Stapleton, and Jonathan Lamb

Before the onset of the COVID-19 pandemic, the Food and Agriculture Organization of the United Nations (FAO)estimated that more than 820 million people in the world did not have enough to eat. The pandemic has only further exacerbated global food insecurity, with at least 135 million people now experiencing food insecurity at crisis levels. In addition to other stressors, such as climate change, natural disasters, and political unrest, the pandemic has thrown into sharp relief the negative impact of global crises on food security. In response, stakeholders across the global food system have sought to mitigate the impacts of these crises, often by using emerging technologies as tools to promote the resiliency of the food supply chain. 

With the support of the Pardee Global Human Progress Initiative and the Tech and Narrative Lab at RAND, we assembled an interdisciplinary team of RAND researchers and PRGS students to examine how selected technologies (blockchain, satellite imaging, artificial intelligence, and the Internet of Things) are used in agriculture and food production and how their adoption may alleviate food security issues in India and Pakistan, as well as across the developing world. In order to understand the use and potential adoption of these technologies, the team needed to conceptualize the global food supply chain—a highly complex, interdependent system. By using mind-mapping, we were able to first conceptualize the socio-technical systems that interact in food production and to trace relationships among the system’s elements. We found that Miro, an online platform, was a useful tool for the project, as it allowed us to collaboratively create visualizations that otherwise might have required in-person brainstorming sessions and analog tools, such as paper charts, sticky notes, and/or whiteboards. Although it should not be the only tool in a research team’s toolkit, Miro allowed us to develop a coherent framework for studying a complex policy problem.

Tools for systems thinking

Emergent complexity—hard to predict patterns that arise over time from smaller interacting parts—underpins most human systems, making them difficult to analyze. Yet, many problems in policy arise due to our inability to anticipate, let alone preempt, issues that result from changes in human systems, including those resulting from human action as well as interactions between system components. As a result, effective decision making and learning in a world of growing complexity requires us to employ systems thinking—the ability to take a holistic worldview “to understand how everything is connected to everything else.” By adopting a systems thinking approach to policy problems, we are better able to produce effective decisions that are also acceptable to a diverse set of stakeholders.  

There are a range of tools that can be employed to better investigate the structure and behavior of complex systems, such as mathematical and computational models and simulations. While these tools can be powerful ways to build knowledge, the examination of human systems also requires qualitative methods to understand the scope of the system and to make informed decisions on how to frame its components and interactions. Mind mapping, for example, provides a visual method for developing a conceptual understanding of socio-technical systems, and it may capture relationships and (inter)dependencies that might otherwise be overlooked.

As a first step in our conceptualization of global food insecurity, we selected mind mapping as a technique to visualize and synthesize the complex landscape of the global food system into a coherent framework. Mind maps are visual representations of ideas that consist of words, images, and colors with the aim of identifying relationships and/or concepts. There are various approaches to performing mind-mapping exercises with collaborators, from the analogue process of putting sticky notes on a board to using digital applications to collect and organize each participant’s ideas.

A variety of digital mind-mapping tools exist, some as browser-based platforms and others as standalone applications. Tool selection depends on a variety of factors, such as the size of the team, tech savviness of team members, research purpose, and intended outputs and audience. We compared several elements of different mind-mapping tools, including each application’s platform type, uses, limitations, and ability to import and/or export files (Table 1). Based on anticipated ease of use, ability to facilitate collaboration, and usefulness for causal loop diagrams, the team selected Miro as our collaborative platform.

Table 1. Comparison of Selected Mind-mapping Digital Tools

Table adapted from The Digital Project Manager’s “Compare The 10 Best Mind Mapping Software Of 2020” and Lifehacker’s “The Best Mind-Mapping Apps of 2019”.

Using Miro to Visualize the Global Food System

The team identified six key steps for completing the mind-mapping exercise:

  1. Inventory system elements (free-list observables).
  2. Add the central theme to a blank canvas.
  3. Elaborate your central theme and identify relevant stakeholders
  4. Brainstorm sub-ideas and related ideas, adding them to the canvas
  5. Use formatting tools to visually organize and display the information
  6. Synthesize findings by bringing the various parts of the diagram(s) together

When we used the Miro mapping tool to visualize the global food system including the selected technologies and various stakeholders, the team’s process unfolded as follows: 

Step 1: Inventory system elements.

Before starting the mind-mapping exercise, we used a brainstorming technique called “freelisting” where each team member created a list of all the observable phenomena they associated with the global food system, including but not limited to actors, processes, activities, decisions, and products. Miro allows for the creation of “frames” to then categorize freelists. For example, we selected frames like “Places and Resources” and “Food Types” (Figure 1). In addition to spurring conversation around potential emergent phenomena, freelists helped the project team to identify boundaries of the system.

Figure 1. Freelists of example system elements


Step 2: Add the central theme to a blank canvas.

Since “food security” was the central theme for the project, we reviewed its various definitions as found in the literature. We selected the United Nations’ definition: “the right to adequate food is realized when every man, woman and child, alone or in community with others, has the physical and economic access at all times to adequate food or means for its procurement.” We also identified a simple diagram of the key determinants of food security from a case study in Sudan to represent our central theme on the Miro board (Figure 2).

Figure 2. Determinants of food security


Step 3: Elaborate your central theme and identify relevant stakeholders.

After having identified the key theme and its determinants, we started our visualization with a macro-level view of the food system and its various components (i.e., production, processing, distribution, consumption, strategic reserves, and import/export) (Figure 3). We used dotted lines to indicate our system’s boundaries. And, we added bolded arrows to show relationships among food system components and light arrows to represent inputs to these components. 

Figure 3. Elaboration of central theme

Next, based on the components visualized for this system, we identified key players and possible decision points for those players (Figure 4).


Description automatically generated

Figure 4. Key players and decisions in the food system

Step 4: Brainstorm sub-ideas and related ideas.

Using the players and decision points from Step 3, the team identified potential pathways and first- and second-order factors that influence decisionmaking (e.g. expected profit calculations and government subsidies influence a producer’s selection of crop types and technologies) (Figure 5).

Figure 5. Key decision pathways

Next, as part of a team brainstorming exercise, we identified various shocks or external stressors (i.e., climate change, recessions, extreme weather events, violent conflicts, and global public health crises) that might steer the system away from a hypothetical equilibrium (i.e. disruption of availability and/or access of food) (Figure 6). 

Figure 6. High-level summary of a climate-related event and its impact on food security determinants

To visualize the impact of climate change as a potential stressor, for example, we considered how potential temperature changes, irregular weather patterns, salinization, and other consequences would impact agricultural yields. In addition, we took into account how those impacts would affect crop prices and incentivize and/or disincentivize the use of certain technologies to mitigate the potential negative impacts. We used directional arrows to indicate positive and negative causal effects of factors involved in the event of a climate-related shock to the system (Figure 7). 

Figure 7. Causal Loop Diagram detailing system-wide impact of climate-related events

We also visualized feedback loops that could potentially trigger behavioral change in terms of crop selection and the adoption of key technologies (Figure 8). 

Figure 8. Behavioral feedback loops for crop selection and technology adoption

Step 5: Use formatting tools to visually organize and display your information.

After developing a holistic picture of the food system and visualizing its key components, we shifted our attention to the role of the four technologies under review to examine how their application, adoption, and use potentially fit in this landscape. We organized this work by using the STEEP framework (social, technological, environmental, economic, political), and color coded different factors that would affect the adoption of these technologies (Figures 9-14). Though we reviewed examples of technology use in agricultural settings across the globe, we had a particular interest in how applications using these technologies might be implemented in India and Pakistan. 

Figure 9. High-level overview of STEEP framework 

Figure 10. Socio-cultural factors identified in the STEEP framework

Figure 11. Technological factors identified in the STEEP framework

Figure 12. Economic factors identified in the STEEP framework

Figure 13. Environmental factors identified in the STEEP framework

Figure 14. Political / Legal factors identified in the STEEP framework

Step 6: Synthesize findings by bringing the various moving parts of the diagram(s) together.

After identifying a range of elements and their interdependencies, the team had a complex and intricate visualization of the global food system (Figure 15).

Figure 15. Snapshots of the Miro board visualization of the global food system

Key Takeaways

Overall, we found that a systems thinking approach was necessary to examine such a complex system, and that mind mapping was a useful technique for visualizing the global food system in order to better understand a seemingly intractable problem—food insecurity. Before applying computational and quantitative tools to investigate complex systems, our work highlighted the value in employing qualitative tools to establish a visualization of the system we were examining. Miro thus provided a useful platform for this exercise by allowing us to systematically assess and synthesize the processes, activities, boundaries, and actors that constitute the global food system. We found that using an online platform for collaboration facilitated the creation and application of frameworks to our research topic, which helped us reach a common understanding of the problem among the team members. It also helped the team to identify various levers to achieve outcomes as we considered the adoption of the selected technologies in agricultural settings. For researchers, clearly articulating the conceptual frameworks of the target policy issue can provide coherence to their research design by informing their research questions, methodologies, and data analysis. We found that the iterative processes of brainstorming, visualizing, organizing, and synthesizing our ideas on an online platform refined the elements of our study.

Despite the positive experience that we had using Miro, we did encounter a few challenges over the course of the project. For one, working asynchronously and remotely were big impediments to collaboration even with an online platform that could capture each team member’s work. The team still required regular meetings to review each individual’s contributions and to discuss and hone ideas and approaches. Additionally, we found that online collaboration was not as seamless as in-person brainstorming due to the differences in our disciplinary backgrounds and experience in visual approaches, which required discussion to ensure that we worked with shared, conceptual definitions and understanding. Still, using a platform like Miro could support a research team’s initial steps to articulate a framework for studying a complex policy problem, which could then be further refined and grounded in evidence by using other methods, such as subject matter expert interviews, literature reviews, and/or in-person focus groups. 

TNL disclaimer: this work represents experimental, exploratory, and often in-progress or preliminary efforts. One goal of the TNL is to get interesting approaches, topics, and concepts discussed and presented quickly. This work has not been peer-reviewed and is not an official RAND publication. No warranties implied or expressed, your mileage may vary, enter as often as you like.

Share your thoughts