From Crisis to Innovation
In April 2023, war broke out in Sudan, triggering one of the world’s largest humanitarian emergencies. Over 11 million people have been displaced—7 million internally and millions more across borders. Camps are overcrowded, aid pipelines are overwhelmed, and the pace of displacement continues to accelerate.
In the midst of this crisis, humanitarian responders face a persistent challenge: delays in delivering aid where and when it’s needed most. Resources exist—but the systems guiding their distribution are reactive, slow, and often based on outdated or incomplete data.
This gap between movement and response has devastating consequences for communities already in crisis. AcaciaPulse was created to address this disconnect—with urgency, technology, and care.
Our Research Roots
Initially, acaciapulse began as a research fellowship supported by the Andrea Chegut Fellowship at MIT in Fall 2024. The project set out to answer one central question:
How can we anticipate displacement and humanitarian need before they unfold—so that response can be faster, more targeted, and more humane?
To answer this, we conducted interviews with humanitarian workers, analyzed displacement datasets, and examined the structural limits of existing tools. The research revealed a clear, recurring issue: existing data systems weren’t just slow—they weren’t designed to respond at the pace or scale of the crisis.
Rather than working with real-time insights, aid actors were often forced to rely on manual reporting, fragmented tracking, and delayed coordination, especially in a country as vast and logistically difficult as Sudan. These insights became the foundation for a different approach—one that prioritized speed, prediction, and usability.
What began as a fellowship project evolved into a working prototype, and later, a live platform supported by the MIT DesignX Accelerator. In building acaciapulse, we collaborated across disciplines—urban planning, data science, UX research, and humanitarian policy—to create a system designed for real-world use.
Methodology
As tweets populate the internet across Sudan, acaciapulse automatically listens for mentions of locations in both Arabic and English
Using machine learning and natural language processing techniques, the model recognizes language that signifies displacement
Using historical migration data, the model predicts where people are most likely to flee given their context. This connects the "pulse" generated on social media to its eventual destination
We see AcaciaPulse as part of a broader shift in humanitarian response—from crisis reaction to crisis anticipation. Our next steps include:
Expanding the platform to other regions affected by conflict and climate displacement
Enhancing model accuracy with new data streams
Partnering with humanitarian agencies to integrate the tool into operational workflows
We believe response can be faster, more just, and more rooted in care—if it’s built on insight, not afterthought.