Navigating the winding routes of data innovation

UNHCR Innovation Service
UNHCR Innovation Service
7 min readFeb 22, 2024

It’s not easy to design and implement cutting-edge data projects in humanitarian settings. That’s why we developed the Data Innovation Roadmap.

Trying to come up with new ways of doing things isn’t a linear process. You do some planning, lay the groundwork, test one approach, discover why it won’t work, take that new information back to the drawing board, try something else. This innovation process is messy and complicated — especially if you’re using emergent technologies or creatively experimenting with data. Such projects “are very overwhelming,” says Rebeca Moreno Jimenez, who runs UNHCR’s Data Innovation Programme and Fund. “They have these additional layers of processing data and building some kind of product using that data.”

The Data Innovation Fund is designed specifically to nurture these projects, providing holistic support to teams from across UNHCR who are creatively and responsibly leveraging the acquisition, storage, processing, analysis, and visualization of data to better protect, empower, and include refugees — sometimes with the help of artificial intelligence or other exponential technologies.

As the Fund was being established in 2022, Rebeca and her team realized that endorsed teams would need very clear signposting to guide them through their convoluted innovation journey. So, they developed the Data Innovation Roadmap.

What is the Data Innovation Roadmap?

Very simply, the Roadmap is just what it sounds like: a guide through the process of data innovation. Using both a map and a text document, it lays out the different stages that projects endorsed through the Data Innovation Fund go through — scoping, research and development, product, and scale — and provides teams with a checklist of the processes required to advance from one stage to the next. It’s a practical tool teams can reference across the 15-month implementation period to check that they’re on track and ready to move ahead.

It’s also a theory of change, laying out what makes a successful data innovation project, and a useful monitoring and evaluation yardstick against which to measure the successes and challenges of each project.

UNHCR’s Data Innovation Roadmap is a guide through the process of innovating with data in humanitarian settings, and a theory of change.

Why is it necessary?

Teams supported through the Data Innovation Fund are working to solve tough challenges at very different scales, in specific contexts, requiring a wide diversity of expertise.

They start out with an aim — for instance, to use data to help UNHCR colleagues make evidence-informed decisions about which communities are most in need of protection. Then, they identify the data they’ll need, how they’ll get it, the ethical or legal or data responsibility issues bound up in that acquisition, and how they’ll process the data and turn it into actionable information. That last bit involves building a product — whether a dashboard, website, or app.

In some cases, the teams include software engineers. Sometimes they don’t. And, whatever their skillset, they’ll be engaging in new — and sometimes confronting — innovation methodologies. Rebeca recalls:

We found that this was overwhelming and very messy. Especially given that this is all extra on top of their ordinary workload. So, the Roadmap helps teams to identify which activities they need to prioritise. It gives them more structure, which helps to ensure they complete their project.

How was it developed?

The Roadmap is an iterative collection of ideas rather than a one-off creation. UNHCR’s Data Innovation team invested time in bringing together relevant resources to inform its design, while learnings from past projects spurred changes and amendments — an evolution that is still ongoing. Its graphic layout was developed by Eirini Malliaraki, a design engineer, who set out to integrate the three different but interconnected strands of work essential to the data innovation process:

  1. Innovation methods — like participatory approaches, co-design, design-thinking workshops, etc.
  2. Data science — data protection impact assessment, legal concerns, research elements, etc.
  3. Data engineering — what hardware and software is needed, etc. This usually kicks off in the development phase, when a team is building their product.

Eirini explains the thinking behind the design:

We aimed to create a resource that is easily navigable, integrating various stages of project development along with checklists, tasks, and deliverables to guide the initial steps of the process without presupposing any prior knowledge.

To enhance self-assessment, we incorporated questions encouraging individuals to reflect on their data’s origins, such as whether it was gathered manually or through automated means; or consider factors such as the data’s velocity in real-world applications, dependence on samples, the array of data types, collection frequency, granularity, and the risk of bias or overspecialization. Our objective was to outline a comprehensive workflow, supplementing it with resources and advice to facilitate the project’s progress.

Our approach focused on understanding the roles, contexts, and pre-existing procedures to develop a versatile tool. It provides a holistic view of the project development process, addressing both the technical and interpersonal elements involved in designing such an initiative.

Previous iterations of the Roadmap, which is evolving as learnings are generated from its application to real-world projects.

How do UNHCR teams use the Roadmap?

Once Eirini had designed the Roadmap, it was up to the Data Innovation team to operationalize it.

They’ve done so by introducing the concept to each new cohort in a kick-off webinar before diving into the initial stage of scoping. “The first meetings are questions, basically,” Rebeca says. “What assumptions are you making? How can you make sure people will buy into the project? Who else needs to be involved?”

Scoping is an essential initial phase of any data innovation project, and many teams will circle back to it later in the journey to check assumptions or redefine their aims based on new information.

Rather than overwhelming endorsed teams by constantly referring to the full Roadmap, Rebe and her team provide that initial overview, then focus in on the different stages as teams reach them. The tool is designed to be flexible — “We didn’t overfit a very prescriptive way of saying ‘this is what you should do’,” Eirini notes — so teams use it in different ways, depending on their needs. They might work on different parts of different stages simultaneously, skip certain steps, or veer away from the outlined approach.

“But whenever they get stuck and ask, ‘What can we do?’, we go back to the Roadmap, and ask ‘Have you considered doing this?’,” Rebeca says. “This is something we have as a backbone, as Innovation Officers, to ensure our teams have a structure to refer to if they’re at risk of getting lost.”

Does the Roadmap make innovation a breeze?

Short answer: Nope! As Rebeca says: “Innovation is inherently chaotic because you’re testing something that hasn’t been tested before.” The Roadmap doesn’t make this journey straightforward, but it does help to identify gaps and guide teams toward the next step in the journey. And often the next step — the way out of the crisis — is to bring more people on board to expand the network of enthusiastic collaborators.

The Roadmap has been effective at “helping teams get unstuck and to advance in a more positive way, so they’re not living in constant chaos.”

Following scoping, teams take sometimes circuitous routes through research and development, and production.

What’s next for the Roadmap?

Like the projects it helps to guide, the Roadmap itself is a pilot. It will evolve as more cohorts use it, delivering important findings about useful amendments or additions.

Reflecting on the rapid advancements within generative AI, Eirini points out the potential need for the Roadmap to evolve, incorporating several adaptations tailored to the diverse array of tools and methodologies employed in different projects. This adaptation would need to account for the unique challenges and processes inherent to various generative AI applications. This may include choosing the most suitable AI models for specific project goals, devising comprehensive data annotation and preparation strategies, developing robust evaluation strategies, and seamlessly integrating GenAI tools with existing workflows of teams on the ground. This raises a bigger point:

With tools like AI, where the frontier is moving so fast, how do you stay timely in the advice you give to your teams?

Pending a broader transformation, the Roadmap has a new look to boost its accessibility. Equipped with this guide, our new cohort of Data Innovation projects — which are tackling everything from proactive disease detection in Malawi to environmental impact assessments in the Darién jungle mapped by Indigenous communities — can venture fearlessly into the high seas of innovation, charting new paths toward an evidence-based humanitarian response.

Explore the new Data Innovation Roadmap and read more about the Data Innovation Fund.

--

--

UNHCR Innovation Service
UNHCR Innovation Service

The UN Refugee Agency's Innovation Service supports new and creative approaches to address the growing humanitarian needs of today and the future.