Acting with Data in Time of Crisis
Unfortunately, we are now living in an era of “permacrisis”, with the covid pandemic combined with the damages of climate change and the dramatic consequences on populations of the war in Europe. In this context, data are essential to provide reliable insights and predictions to support the different stakeholders involved in the response to these crises, and to better help the affected people. While some constraints can be expected by working on these types of issues, such as the need for rapid development and delivery cycles or the limited availability and turnover of the NGOs teams and volunteers on the ground, there are some well-known best practices in conducting data projects that help meet the challenges in times of crisis.
Responding to these crises is a collective societal challenge, so the more we share, the better. In this post, Axionable’s data experts share some concrete lessons for the data practitioners’ community, drawn from our experience working with the French Red Cross (Croix-Rouge française) to support their actions for Ukrainian refugees.
The imperative of an innovative & hybrid business model
To act in times of crisis, actors from different backgrounds, from the public and private sectors, must join forces and implement innovative operational models to face the crisis. We are convinced that pro bono (i.e. having data experts work for free on a project) is a powerful lever to act in a very short time, and bring the necessary expertise that is sometimes lacking. However, it might be beneficial for all to imagine a more sustainable model in the medium and long term that combines non-profit and for (responsible) profit companies (B Corp companies for example).
Data culture & the sense of emergency
We are fortunate to work with the French Red Cross (la Croix-Rouge française) staff and volunteers involved in the NGO’s response to the Ukrainian crisis. We find it very inspiring and personally admire their commitment and dedication to causes such as refugee assistance.
Volunteers don’t think about “data-driven use cases to maximize refugee satisfaction” as data consultants might say, but they have concrete challenges and operational burdens when helping Ukrainian refugees with their basic needs: food, drink, shelter, medical care, psychological support, etc. Data can help in solving these pain points. As always, start by the problem, not the data !
The sense of priority and emergency is a tremendous catalyst for user adoption. Assuming data can rapidly solve some operational burdens, there is a perfect alignment between data users and data developers.
Tailoring data project organization when working with volunteers & NGO
NGOs volunteers come from a variety of personal and professional backgrounds. This needs to be taken into account when onboarding project stakeholders to ensure that everyone is on the same page.
In terms of project organization, working within a diverse ecosystem of stakeholders underscores the imperative for a clear data governance model, that is explained in a pedagogic manner. Most of the volunteers don’t know what the acronym RACI stands for, but they clearly understand when it is explained in terms of concrete tasks, responsibilities and sharing information on the ground.
Moreover, volunteers often have a high attrition rate within NGOs. This aspect emphasizes the need for a constantly updated project & code documentation in order to ease the onboarding of new volunteers. This mainly concerns the data architecture, data dictionary, data pipelines specifications and code commenting.
Finally, the agile approach is very suitable to act in times of crisis. Being able to demo and release new features every 2 weeks is relevant to meet urgent quickly evolving needs while gaining the trust of users by providing concrete solutions quickly. Be careful not to overwhelm the project team with agile jargon such as spike, time-box, etc. Keep it simple.
For all these reasons, we created our own method called ADM for Data (Axionable Delivery Method) which consists of modeling, within an Atlassian Confluence space, all of the templates required for project management, data product documentation, or GDPR compliance. Our method has proven to be useful and a real booster to act in times of crisis.
Quick and dirty resilient data products
Acting in times of crisis means being under pressure to quickly deploy data products into production. There is no room for R&D. Therefore, there is also no room to compromise on privacy, cybersecurity or resilience when working with Ukrainian refugee data.
We used Google Cloud Platform ecosystem as a booster, combined with Big Query for storage / query, Data Studio for dataviz, Terraform for Infrastructure as a Code, Cloud Functions for data processing, GitLab for code versioning, Secret Manager and Cloud IAM services for security and the whole Google Cloud security stack (Cloud IAM, Secret Manager, etc.) as a solid foundation to build on top of.
This data architecture might seem “overkill” to make dashboards, but in practice it has proven necessary. As an example, Terraform came in handy when we needed to change our GCP cluster instance because of internal constraints. While the redeployment only took us a few minutes, it would have taken us days without an IaaS service such as Terraform.
In addition, it was necessary to ensure that the solution could be maintained over time by the Red Cross itself. Indeed, the staff and volunteers should be able to take over the solution, maintain it and make it evolve beyond our one-time intervention. So, it was certainly not the right time to experiment with risky or exotic technology choices.
In times of crisis, the robustness and resilience of the data infrastructure is a must have. As crises are likely to multiply, it is crucial to be able to reuse the same organizations, methods and technologies each time in order to provide a rapid response. It is therefore essential to build reusable data products.
The “permacrisis” era
The AR6 IPPC report states once again how living conditions will evolve drastically as a result of climate change. Climate-focused NGOs such as Finance Reclaim, WWF or GreenPeace are acting as whistleblowers by providing data and insights. Open data platform such as Copernicus help to better anticipate and adapt to these challenges.
Data has had a proven role in the fight against the pandemic crisis, as shown by the case of Covid Tracker, where Guillaume Rozier successfully set up a team of volunteers to develop this tool. APHP also setup a team of volunteers to help them on handling their data during the covid crisis.
It is therefore certain that data have a lasting role to play in times of permacrisis.
A big thank you to the contributors of this post : Enora Georgeault, Gaël Charles, Paul Laville, Mickaël Nicolaccini and Imène Boumghar
