Interview with Martin Noblecourt, from CartONG
Six years after our first interview in 2018 with CartONG, an NGO set up in 2006, Défis Humanitaires takes another look at the issues surrounding Geographic Information Systems (GIS), new information and communication technologies (NICT) and, above all, data management for humanitarian and development action. With Martin Noblecourt, Senior Research and Fundraising Officer at CartONG:
- Hello Martin. On 8 October, CartONG published the updated version of its study on the challenges and needs of NGOs in terms of data management tools. You are a co-author of the study; before going into more detail, could you remind us what CartONG’s expertise is, and for whom?
CartONG is an association set up by cartographers in 2006: initially specialising in GIS, our core expertise has continued to expand, and we now work on the entire data cycle and information management tools. We have around forty employees, mainly in France, and a network of volunteers, enabling us to support a wide range of organisations, whatever their resources.
We work with a wide variety of players in the international solidarity sector, including major international NGOs such as Médecins Sans Frontières, Solidarités International and Terre des Hommes, United Nations agencies (UNHCR, UNICEF), as well as smaller French and international NGOs (Dahari, Asmae, Sidaction, etc.). CartONG is also involved in sharing networks between NGOs, in conjunction with other ‘support’ NGOs such as Groupe URD, the Bioforce Institute and our peers in the H2H network internationally, and of course the OpenStreetMap free mapping community.
We provide direct support to organisations through training, implementation (e.g. data collection, mapping, development of tools such as dashboards, etc.) and strategic advice, but we also act as a think tank for the sector in general, thanks in particular to funding from donors.
It is in this context that for several years we have received co-funding from the Agence Française de Développement to support French-speaking NGOs in managing their programme data. The second phase of this project began in 2024, and we wanted to update the report written in 2020 on the situation and challenges of data in international solidarity. To do this, we carried out a survey (54 organisations responded), interviews with specialists (18), workshops (78 participants from 23 organisations) and a literature review (over 100 sources).
- In 2020, in the first study, you referred to programme data as ‘the new Eldorado of international solidarity’… Can you expand on this notion of ‘programme data’, and tell us whether this ‘Eldorado’ has been confirmed? You note that, over the period 2013-2025, there has been an ‘explosion in the volume of data produced’, which has increased 40-fold! But you also point out that there is a question mark over the quality of this data, and a lack of evidence as to its positive impact on action on the ground…
In this new report entitled ‘ Beyond the numbers: reconciling innovation, ethics and impact ’, we set out to see whether the promises of the ‘data revolution’ had been fulfilled. Four years on, we find that it is still difficult to find solid evidence of the impact of using data to guide organisations’ choices. Data collection is still seen primarily as necessary for bottom-up accountability, particularly to donors, rather than for operational use in the field.
Generally speaking, digital and innovation projects are rarely evaluated (ALNAP found in a review of 540 humanitarian innovation projects that only 16% had any evidence of impact). Many of the data-related tools that have been heralded as technological revolutions in recent years, such as drones, blockchain and, more recently, AI, have certainly been successful in certain sectors but have not transformed the work of humanitarians because of their limitations.
We connect these challenges with the recurring problem of data quality, a real ‘eternal quest’! There are still many problems of bias and under-representation – for example, in terms of language, as most surveys are not designed in local languages – not to mention recurring problems with the quality of technical methods and tools in the field. These concrete, human aspects cannot be magically resolved by technology, even if it can help.
These underlying problems are reflected in the under-use of qualitative data (seen as less robust) and secondary data (i.e. the re-use of existing data, which comes up against a lack of a culture of sharing) – even though AI could have a positive effect on the latter by simplifying data sharing, and there are promising initiatives in this direction.
- In the new study, you note progress in the appropriation of this concept of ‘programme data’, and the development of skills and specialisation. Data literacy, which you see as a crucial factor in ‘getting everyone on board’ internally, has improved… What’s your analysis?
The data management sector is indeed continuing to professionalise – and we’re delighted to be playing our part. We can see that the concept of ‘programme data’, introduced by CartONG in 2020, is beginning to percolate. Skills are developing, even if the ‘basic’ tools (Excel, ODK/KoBo, SurveyCTO, ArcGIS, QGIS, etc.) are still the most widely used, particularly by local organisations.
In addition to individual skills, we are also seeing an increase in capacity at organisational level, with greater specialisation in data-related professions. For example, while less than 1% of respondents to our 2020 survey felt that the data culture within their organisation was ‘complete’ or ‘sufficient’, this year the figure has risen to almost 30%. However, getting all organisations, and in particular decision-makers, to realise the need for a global approach to data – which goes beyond simply ordering a dashboard – remains a challenge.
- Beyond the rise of Geographic Information Systems (GIS), you observe that it’s basic tools that remain the most widely employed, particularly by local organizations, and that there are big differences depending on the sectors and areas of intervention. Could you elaborate?
It’s a recurring observation when we go out to support our partners in the field, and one that the survey confirms: for many international solidarity organizations, whether local or international, recent tools such as biometrics, voice recognition, AI, big data, etc. remain inaccessible. Most data collection is still carried out on cell phones, or even on paper, and information is mainly analyzed on spreadsheets, whether online or offline (such as Excel or Google Sheet). In fact, we have noted a boom in the use of GIS, particularly among national organizations, perhaps reflecting the intermediary status of these tools, which are more accessible than statistical software, but provide powerful visualization that can be understood by everyone.
Beyond this general observation, there are major disparities between organizations, regions and sectors. Firstly, between international and local NGOs: even if local and national NGOs are making efforts to equip themselves with tools, the IT systems and Internet connections available to them remain limited, and they still often rely on paper or offline tools. Then, between regions: socio-economic and educational conditions, as well as the existence or absence of a private sector, generate major differences in skills in different areas, with a concentration of projects and skills in East Africa, South Asia and the Middle East. Finally, within organizations themselves, certain sectors (EHA, mine clearance, cash transfers, etc.) are more amenable to digitization than others.
- The study observes that there is a profusion of technical solutions available today… but that it remains difficult to identify the right ones… Similarly, there is still a lack of funding for the definition of needs and sustainable appropriation… What is your analysis on these points?
It’s also one of CartONG’s long-standing hobbyhorses: as the NGOs say in our survey, what they need today is methodological support, support for structuring, training, and therefore funding for this, much more than inventing yet another new AI app or tool. There’s nothing revolutionary in saying this, but our colleagues working on localization are saying the same thing: if we don’t provide organizational support, innovation won’t be able to scale up.
CartONG is working on this through our training offer, our learning corner with royalty-free tutorials, and many other NGOs are moving in the same direction, but it’s not enough. While there is now a real appetite among NGO staff for the issue of data, self-learning is not enough, and it’s no substitute for initial and ongoing training, or for having a strategy for increasing skills on the subject.
- CartONG warns of what you call “bottom-up accountability” in the collection and use of these data, i.e., NGOs tend to use them more in terms of accountability to donors than to beneficiaries. You also speak of a risk of pressure towards a quantitative and globalized approach to data collection, versus an approach that is more attentive to adjusting data to the specific needs of local populations. Finally, you mention a risk of “techno-solutionism” in the humanitarian or development approach… Can you elaborate on these points?
NGOs (and other research has made the same observation in the public sphere at the National Statistical Institutes of Southern countries) tell us that they collect data primarily because donors ask them to. Whatever one thinks of accountability to funders, it’s not right that accountability to the populations whose data is collected should not be a priority.
And when NGOs do develop accountability mechanisms for affected populations, they are reactive (complaint or feedback mechanisms) and not proactive. Projects, and therefore the data that will be collected, are still rarely designed in conjunction with local populations – we cite the work of the NGO Ground Truth Solutions, which helps humanitarian organizations change their methods in this direction. This would require a profound reform of project management in the humanitarian sector (as the F3E is doing, for example, with change-oriented approaches) and of our relationship to transparency.
Otherwise, we run the risk of falling into techno-solutionism, imagining that by adding more and more technology we’ll be able to get around these system design problems. To give an example, a specialist told us that international players had asked him to try to use AI to extrapolate statistics (on the needs of populations) for a country in crisis from data in neighboring countries, because it was too complicated to go out into the field and collect them… you don’t need to be an expert in the human sciences to understand that this is nonsense.
Eventually, we’ll end up with a system where everything is counted using ever more quantitative data, and endlessly ground up by AIs, without any qualitative approach or consideration of the context and origin of this data (the famous “metadata”).
- You also speak of the risk of a “two-speed humanitarian system” or a “digital divide” when it comes to data management. What do you mean by this?
When we wrote this study, we had the feeling that the humanitarian system would end up operating at 2 speeds if nothing was done, and we wanted to draw attention to this. On the one hand, we have the United Nations, the major international NGOs, donors and institutions in Northern countries, which are rather well endowed (financially and above all in terms of skills), can appropriate innovations to further improve their operations, can hold their own against the private sector, and have high expectations in terms of data usage. On the other hand, the rest of the ecosystem – most NGOs (local NGOs in the South, but also “medium-sized” NGOs in the North, i.e. most French NGOs!) and public authorities in the South – are struggling to retain their talents, are subject to innovation and the law of the “GAFAMs” (big tech companies), and are “forced” producers of data that they struggle to exploit.
We cite a case study from a Somali consultancy (Somali Public Agenda), which paints a vitriolic picture of the hierarchical division of data labor in their country, with design, supervision and analysis solely in the hands of international experts, and local researchers confined to collecting data in the field and never seeing the results of their work – with all that this implies in terms of the impossibility of capitalizing locally… The opposite of the localization and decolonization of aid, then!
This reflects a vicious circle of “forgotten crises”, where a lack of local resources and skills leads to a lack of data, and consequently to insufficient allocation of funds, which only makes the situation worse…
- Two concepts are close to your heart: localization issues with regard to data governance, and responsible data management, a point on which you note progress, but stress that it is still difficult for some players to take into account… What are we to understand? Moreover, you cite the examples of “beneficiary screening” and the suggestion by the US Information Technology Industry Council (the GAFAM lobby) to stop aid to countries “whose actions run counter to US techno-economic interests”…
Localization (and behind it the question of colonialism and governance) is a central, one might say moral and ethical, issue for the international solidarity sector, and we wanted to show (after an initial work on the subject – “ Changing perspective: for a local approach to data ” – in early 2024) that it was also a central issue for data. In fact, that’s what the respondents to our survey felt, placing it as the No. 1 challenge for the coming years. This encompasses “classic” localization challenges, such as the empowerment of local players, the difficulty for international NGOs to support their local partners, and for local NGOs to obtain funding, etc.
But the data sector, and the digital sector in general, also presents specific challenges. We can illustrate this with two examples: firstly, the screening of beneficiaries, which, I would remind you, led to Coordination SUD and other French NGOs launching the biggest protest against the Agence Française de Développement in recent years, all the way to the Conseil d’Etat, where they won – on a data issue, in other words. This is what we call “excessive compliance”, when the demands of donors are incompatible with the principles of responsible data management, and therefore by extension humanitarian principles, and there is no longer any proportionality between means and ends, forgetting the rights of the populations concerned.
International and local NGOs today find themselves at the center of an increasingly complex “geopolitics of data”, between the standards of their home governments (such as the RGPD), governments in the South who are developing their digital sovereignty and their own standards (rightly so!), and of course the big tech players, or the GAFAMs as they would say in France.
As far as our relationship with the latter is concerned, we’re witnessing fundamental movements (such as Microsoft’s return to prominence in NGOs with its Office365/Azure systems in recent years) that are not being thought through, discussed and politically arbitrated as to their consequences. We quote from a memo from the NGO Development Initiatives, which actually unearthed a memo from the GAFAM lobby explicitly recommending that the US government stop helping countries that wish to develop their digital sovereignty (data localization)! Today more than ever, the digital tools used by NGOs are not neutral.
- The study refers to the notion of “sustainable data”; can you elaborate on this?
CartONG identifies 4 complementary components: firstly, sustainability itself, i.e. not only the ecological sustainability of systems and the sobriety of equipment, but also the fact that they are resilient. Secondly, responsible, protected data that is secure against cyber-attacks and of high quality. Then there’s inclusiveness, taking into account the fractures already noted between countries, genders, languages or even people with disabilities. Finally, sovereignty, with sustainable business models that do not generate excessive dependence on private players.
Today, the same thinking is found within the French administration and local authorities. Alliances are therefore possible to build a different kind of digital.
- You note that the question of the future of Artificial Intelligence for humanitarian and development NGOs remains… That future, you say, depends on quality data and good practice. Can you elaborate on this AI challenge for our sector?
Beyond consumer applications, and behind the messianic rhetoric of certain entrepreneurs, the question of AI does indeed refer to far more prosaic issues of data quality and digital acculturation. As one specialist told us, if we’re not careful, we’ll be importing our bad habits and data biases into AI, but this time with an impact multiplied by the power of the tools! AI, based on a corpus of existing data that is essentially Western, will reinforce existing power dynamics if nothing is done. And the pro bono/“for good” initiatives of major tech companies are often aimed more at supplementing their corpus of data (e.g. for advertising purposes) where they have little of it, rather than actually serving a humanitarian cause.
To sum up, AI presents many challenges: data bias, model drift and algorithmic errors that are harder to detect with massification, ethical considerations, reinforced disconnect with communities… Faced with this, we need to return to simple good practices: start by experimenting, be wary of GAFAM’s fine promises, understand how data is processed, implement risk management, stay focused on real problems to be addressed rather than doing AI for AI’s sake, or be transparent with communities. There will be uses that will develop quickly and easily (for example, tedious data cleansing tasks), but we’ll need to be more vigilant when we’re dealing with personal and sensitive data, or with relationships with the populations affected.
- Today in 2024, what technical tools are available in terms of Geographic Information Systems (GIS), information and communication technologies (ICT) for NGOs… and training? More specifically, can you describe the “toolboxes” offered by CartONG?
The tool base hasn’t changed that much: behind the essentials (GIS software, spreadsheets and database systems, forms tools, satellite imagery…), we’re seeing a rise in tools often focused on simplifying processes and data visualization (online cartography/webmapping, dashboards/dashboards). This is consistent with our aim of spreading data culture, as long as neophyte users retain a modicum of critical awareness of the data presented to them, which is more accessible via these simplified tools.
In the same vein, ongoing experiments to facilitate data sharing through AI (Deep.io – UN consortium, IFRC and service providers; GANNET – Data Friendly Space; or SOPHIA – ACAPS) or to improve machine translation into local languages (ClearGlobal) take advantage of technological progress to improve the system.
CartONG continues both to analyze these developments (for example, we are currently running a research-action project with several partners on AI, or a project on anticipatory action and localization as seen through data) and to produce resources that can be easily used by all players, whatever their size. The entry point is the “ learning corner ” where we centralize toolboxes and more specific training courses: to discover data management, using Excel, GIS, mobile data collection, OpenStreetMap, or even responsible data management, data visualization, qualitative analysis, etc. Early next year, we will also be testing an open training course for local and international NGOs in Dakar (with the support of AFD), to work on this acculturation as close to the field as possible.
- Do you have any examples of how these tools have been put to practical use, and how they have been evaluated, in current emergency humanitarian situations?
As I said, we do very little impact assessment on the data tools themselves… the shoemakers are the worst shod! We do, of course, have numerous case studies on various CartONG and partner projects, but few make it possible to clearly distinguish the specific impact of data. As a manager from another H2H organization put it, “people absorb data and information, but they don’t always consciously remember what they used at any given time and how it influenced their action or decision”. On the other hand, program data management is essential for measuring the impact of humanitarian projects (it’s a key component of monitoring-evaluation) and improving this measurement.
We tend to have thematic evaluations, such as the impact of digital cash distribution tools, the impact of anticipatory action (which relies on data), or the impact of GIS in the field, as assessed by MSF, for example. There are also numerous capitalizations/learnings on projects, such as the one we produced following our collaboration with the RESILAC consortium in Lake Chad (ACF/Care/Groupe URD).
With regard to current crises, many major humanitarian organizations (such as our partners MSF, UNICEF, UNHCR, etc.) have well integrated data management functions, with services (mapping, needs assessment surveys, operations dashboards, etc.) that are now indispensable to the day-to-day work of their teams. The challenges we have addressed in our study (localization and accountability to populations, cybersecurity and data protection, integration into organizational strategy, generalization of data culture, standardization and sharing, etc.) are now more at stake than demonstrating the usefulness of these technologies… even if, as we point out in the study, we lack real impact studies of data – and digital in general – in our sector. To do this, we need to have the resources to carry out this type of study, which is almost never the case !
- Do you have any concrete examples in mind of the impact of these tools on the “humanitarian-development Nexus”?
Groupe URD is one of our closest partners, so CartONG has been trying to project itself into this Nexus logic for a long time, notably through our approach to data localization. We’ve noticed that examples of innovative and impactful data tools generally come from countries where the digital and data ecosystem was already mature before a crisis (e.g. Lebanon, Nepal, Kenya…). In countries where such ecosystems do not exist, humanitarian actors will need to think about how to avoid interfering with the development of these ecosystems through their interventions (whether by “stealing” talent or by building their own systems parallel to public statistics), and make the link with development actors so that they take an interest in the value of their data in the event of a crisis.
Anticipatory action seems to us to be a good example, reconciling the emergency approach (we anticipate the distribution of aid) and the long-term vision (we build trigger indicators, but also the type of response, with local actors and communities).
- In conclusion, what message would you like to send to your partners, to humanitarian and development NGOs, and to our readers?
The program data sector, like digital technology in general, is constantly evolving. This speed, coupled with the multiple challenges we identify in our study, can seem daunting. But there are also some great success stories that we’re highlighting! The most important thing is for NGOs, and all humanitarian actors, to continue to invest in the subject, and to take a critical and lucid look at it. It would, of course, help if donors were to provide a little more support for their initiatives and structuring. In fact, the ALNAP study I mentioned on humanitarian innovation shows encouraging results, despite the lack of funding for scaling up, and the fact that our sector’s innovation budget is proportionately lower than that of the paper industry, for example…
And we encourage players to take part in discussion forums on this subject, such as the exchange days organized by CartONG: together, we can do a lot! We will continue to play this role of catalyst and link, remaining true to our mission, which is to enable NGOs to be more effective in helping communities affected by crises and building their resilience.
|
Camille Brunet, Head of SERA (Monitoring, Evaluation, Accountability and Learning) & Quality Approaches at SOLIDARITES INTERNATIONAL |
| “In recent years, more and more data has been generated by our operational activities, mainly to serve upward accountability, requiring, among other things, better structuring. Solidarités International (SI) thus bears responsibility for the safe and responsible management of its data, particularly its program data, also and above all for downward accountability, to the people helped… This positioning is essential to ensure quality steering of our activities and accountability to the various stakeholders.CartONG’s support began as early as 2015 and a partnership agreement has been established in 2019, to support our teams on these methodological, technical and ethical issues around program data. Most recently, CartONG supported our teams in Syria in improving their data management, with two priority objectives: to rationalize and centralize data, and to create an interactive map to ensure a better flow of information, and strengthen coordination and responsiveness… to ultimately maximize our humanitarian impact”.
|
Interview by Pierre Brunet
Writer and humanitarian
Martin Noblecourt
After training in social sciences and international solidarity management & policy, Martin Noblecourt joined CartONG in 2012, first as an administrative officer, then as project manager.
In 2016, he initiated CartONG’s involvement in the international Missing Maps mapping project and has since been an advocate of OpenStreetMap and collaborative approaches both internally and externally. He has managed numerous participatory mapping and open data projects, with a strong interest in building partnerships particularly with local organizations, and promoting new approaches (sensitive mapping, open data collaborations, etc.). Today, he puts his experience at the service of CartONG’s fundraising and the organization’s development, and contributes to its research and capitalization activities. Outside CartONG, Martin is a local politician, first deputy mayor of Chambéry.
I invite you to read these interviews and articles published in the edition :
- Towards a contagious humanitarian “domino effect” ?
- Lebanon, Gaza, Iran, and now what ? A personal interview with Antoine Basbous, director of the Observatoire du Monde Arabe.
- On the humanitarian front line in Ukraine. Interview with Mathieu Nabot, Country Director of Solidarités International
- Summary: “Falling short ? Humanitarian funding and reform”
- Solidarity with Armenian students.