Putting a little-known route on the map: Developing baseline data from scratch

It was 2021 when UNHCR, the UN Refugee Agency, heard about a boat of migrants and asylum-seekers washing up on a beach in Comoros, an island nation hundreds of kilometres off Mozambique’s eastern coastline.  
 

Until then, movements along this part of the Western Indian Ocean route had been largely associated with Comorians attempting to reach the island of Mayotte, as well as a separate route used by Malagasy departing Madagascar. “It was the first time that we became aware that third-country nationals, in this case, Congolese, Rwandans, and Burundians, were making this journey through the Western Indian Ocean,” recalls Guillaume Potie, then Associate RSD Officer with UNHCR's South Africa Multi-Country Office.
 

This report caught UNHCR’s attention: it suggested a larger, mixed flow of people travelling along this maritime route, aiming – most likely – for Mayotte, a French overseas department south-east of Comoros. Except for sporadic Francophone reports about drownings and strandings, information about this route and its risks was thin on the ground. Referring to these reports and accounts from people on the ground, Guillaume observed: “From what we heard, a similar number of people could have been dying along this route as on other deadly migration routes, yet this route has been largely overlooked.” 
 

This kind of underexposure has real-world consequences. If a route isn't documented, people can disappear without being counted, abuses remain hidden, and there is little basis or incentive for changing policies or deploying assistance. Guillaume and a team of colleagues set out to address this knowledge gap by establishing a baseline dataset, on which further research – as well as advocacy and programming – could be built.  
 

With support from UNHCR Innovation and researchers from the Mixed Migration Centre (MMC), they’ve done just that.  
 

A panoramic response to human mobility 

In recent years, there has been an increasing acknowledgement that refugees and asylum-seekers often travel as part of mixed movements, in which people with different profiles and reasons for moving use the same routes and means of transport, often in irregular and precarious conditions. During these journeys, people face shipwrecks, drowning, abuse by smugglers, exploitation, detention, violence, family separation and barriers to seeking asylum. UNHCR’s route-based approach is a means of “responding more effectively to the challenges of mixed movements.” It takes a panoramic view, attempting to keep people safe along the way by engaging with stakeholders across the entire route – in departure, transit, and destination countries.  
 

Whereas mixed-movement routes across the Mediterranean are comparatively better documented and studied, information on movements by third-country nationals through the Western Indian Ocean remains much more limited. This lack of information was, Guillaume and his colleagues realized, a critical concern. To take joint action to save lives, we need to know where the route is, how it’s being used, and its risks. 

Framing the challenge 

The 2021 report suggested that people fleeing conflict-affected countries in mainland Africa were now turning to this route from Comoros to Mayotte. “We quickly realised this wasn't going to be a one-off. It was a recurring movement that we needed to understand and prepare for” recalls Guillaume. 
 

According to UNHCR’s records, around 100 third-country nationals aiming to reach Mayotte, got stranded in Comoros that first year. In the years that followed, the number of cases recorded increased. Gaps in refugee protection frameworks in some parts of the region alongside data deficiencies made it challenging to determine the scale of the issue. However, the available information pointed to significant protection risks. 
 

A French Senate report suggested up to 10,000 people – primarily Comorians trying to reach Mayotte – could have died in the sea between the two islands from 1995 to 2012. One Comorian official, in 2015, informally used the figure of as many as 50,000. (There are no more recent estimates.) 
 

While reliable data on third-country nationals was virtually non-existent, accounts gathered by UNHCR during missions to Comoros suggested they too were dying at sea, including testimony from two survivors who said they were the only people to survive after a vessel carrying more than 20 passengers capsized. In addition, people on the move were regularly being detained in makeshift facilities in Comoros (which does not have an asylum framework and is not a signatory to the 1951 Refugee Convention), often in poor conditions. 
 

What was clear was that the route was underreported and protection needs were high. Gaining a better understanding of these mixed movements would require cooperation from a complex array of actors from both within and beyond UNHCR – given that the route likely spanned at least three regions – and delicate navigation of political sensitivities. Guillaume and his team applied for the Data Innovation Fund in 2024 with the challenge, but without a clear concept of how to address it. 

Starting from scratch 

Initially, the team tried to find a specialized, innovative data collection approach. But to use artificial intelligence, for instance, to develop predictive analytics, baseline data was a prerequisite. In the end, the team determined that developing a research methodology to simply establish baseline data in this new, evolving terrain would be innovative enough. Partnering with MMC, a knowledge hub within the Danish Refugee Council, brought onboard researchers Ayla Bonfiglio and Jim van Moorsel.  

In recalling the point of departure for MMC’s collaboration with UNHCR, Ayla explains, “Every now and then we see alarming news pieces about capsizing boats, about people taking difficult journeys for upwards of a month at sea. And yet we realized that there were no good estimates about the number of people on the move, no regular data collection.” 
 

This underexposure was likely due not only to the fact that this was an unexpected route for third-country movements, but also that the territories involved did not host highly active civil societies motivated to draw more attention to the challenges. “For France,” notes Safa Ezzeddine, Associate Statistics and Data Analysis Officer with UNHCR France, “the study was also very important because there were very limited data on refugees and migrants in Mayotte – their profiles, their routes, their intention, their vulnerabilities, and their protection risks.”  

Reversing the research 

The maritime and branching nature of this route made data collection highly challenging. “Arrivals could be on one coast or another, and that makes monitoring these movements, and equally responding to people’s needs, extremely difficult,” Ayla says. As an under-prioritized route, resources to underwrite on-the-ground research at different points were scarce; support from UNHCR Innovation was a rare chance to conduct that research.  
 

On another mixed-movement route with big numbers, a researcher might start with a quantitative review to capture trends before zooming in on qualitative research. In this case, that approach was inverted. “We were forced to start with in-depth qualitative work to identify cases of people travelling along this route,” Ayla says, “to ask them longitudinal questions, to understand what they experienced along the route and what their knowledge was of the route going forward, and then build out the picture.” 

Using information from conversations with experts, stakeholders, and people on the move, MMC then built a quantitative research design that seemed feasible. “But it was complicated by the fact that this region is also very climate-vulnerable,” Ayla notes. “And when a climate disaster strikes, the capacity of actors on the ground to participate becomes challenging.” A cyclone in Mayotte significantly slowed data collection, disrupting the work of the local partners conducting the surveys. Since people move fast through different jurisdictions, working with such groups – who have a long-lasting presence on the ground – was essential. 
 

Ultimately, 22 qualitative interviews and three focus groups discussions were conducted in person with refugees, migrants, and key stakeholders in Madagascar, followed by a quantitative surveying phase that yielded 513 completed surveys in Madagascar, 22 in Comoros, and 50 in Mayotte.  

Uncovering important insights 

After two years, UNHCR and MMC had collected enough information to dramatically expand understanding of route dynamics. “In the first year, we managed to map the route and understand what is happening,” says Bob Beerens, who took over Guillaume’s portfolio in 2025, “and that should be sufficient information to bring the right people to the table.” A briefing note, an in-depth report on Madagascar, and infographics were produced to disseminate the relevant data.  
 

The research confirmed certain assumptions and revealed some unexpected realities. As expected, people in transit on this route are moving due to different motivations and with different final destinations in mind. In addition to the countries assumed to be involved, Madagascar was also found to be a transit country, adding a new branch to the route map. Some Somalis were departing by sea from Somalia, undergoing a one-month maritime journey.  
 

One key finding was the significant role smugglers were playing not only in facilitating movement but in defining the desired destination – particularly for people who were already on the move and unsure of their next steps. “This is a big departure from what we see on other routes,” Ayla says. “For some people, what stands out here is that smugglers are actually shaping the destination choice.”  

The big picture 

Although many quantitative questions remain, information about who is moving, the drivers of onward movement, and protection concerns along the route provide a solid basis for advocacy. 
 

In December 2024, the project team presented their findings, validated them with key stakeholders, and made recommendations for action. This gained significant visibility and has supported advocacy interventions in relevant countries. “What we managed in the last year and a half is to put the Western Indian Ocean route on that map, which is already a massive gain,” says Bob. 
 

As a basis for conversations – and for getting academics interested in developing further data, which is sorely needed – the materials developed through this initiative have been invaluable. Time, Ayla says, is now the most important factor in collecting more data – given how long researchers need to spend on the ground to encounter people on the move and build trusted relationships. 
 

“It is definitely the earliest stage we’ve ever entered a mixed movement context,” she reflects. “So, it will be also interesting for us to trace the impact of this work with that in mind.” 
 

For now, the project has helped improve recognition and mapping of these movements and has established a stronger evidence base for governments and humanitarian actors. UNHCR is better able to advocate for measures that will protect human lives and dignity – and a foundation has been established on which further research, monitoring, and programming can be built.  

Find out more about our Data Innovation work here.