From Prototype to Impact

What began as a machine learning model to predict floods and droughts has evolved into a broader tool for climate resilience. By monitoring river reservoir levels, the system aims to forecast hydrological risks. 

User engagement is consistent, with agriculture-related content making up 50% of platform activity—underscoring its relevance for farmers. 

One user shared after the October 2024 floods:

Expanding Features and Shifting Models

What began as an early warning tool has since evolved into plans for a broader climate resilience hub. Future platform developments will support post-hazard response planning, including health and safety updates, access to aid services, and farm recovery support. 

“We noticed a significant uptick in user participation after floods, especially from households looking for post-disaster information,” said Sikem. 

On the business side, Map & Rank is shifting from a government-centric model to a more agile B2B approach. They are actively engaging farm input suppliers, microfinance services, and crop insurance providers to join the platform—offering tailored analytics and risk data. 

The team projects a 40% reduction in economic losses for registered users in future flood cycles and aims to preserve 300,000 hectares of floodplain by promoting resilient crops like cassava and cashew. 

Navigating Data Challenges in a Complex Environment

Building a climate intelligence platform in a data-scarce environment is no small feat. For Map & Rank, one of the biggest hurdles was accessing the right data at the right time. 

“We need different types of data—geospatial, ground-truth, and community-sourced,” said Sikem. “While we could control drone mapping for geospatial data, getting access to government-held water discharge measurements took months.” 

The team also faced frequent power outages, which disrupted data processing and model training. To address this, they plan to relocate to a facility with backup power and are establishing early partnerships with national data agencies like Cameroon’s Climate Change Observatory. 

With many watersheds lacking water gauges, the team relied on community reports to validate their machine learning models—an approach that, while innovative, introduced variability in accuracy. Despite these obstacles, Map & Rank achieved a 60% accuracy rate and is actively working to improve it. 

Open Source, Frontier Tech, and the Road Ahead

Map & Rank has embraced open source to accelerate development. They’ve established CI/CD pipelines, published developer documentation, and opened their GitHub for public contribution. 

“Open source gave us structure and access to external expertise,” said Sikem. “Balancing performance and compatibility—especially with our mapping tools—was tough, but we’ve made steady progress.” 

Despite infrastructure gaps, their AI models continue to scale, supported by community-validated data and growing technical capability. 

Looking Forward

Map & Rank’s ambitions stretch beyond early warnings. They are now exploring predictive analytics for crop yields, aiming to help farmers adapt planting strategies based on forecasted climate risks. 

“Our solution can equally be leveraged to measure and predict crop yield based on climate risks forecast and suggest optimization options for farmers in vulnerable communities,” the team shared. 

The team’s mission is rooted in community resilience. From day one, they’ve focused on building trust, enabling action, and closing the loop between local insight and institutional response. 

A Year of Growth, Mentorship, and Momentum

Map & Rank’s time with the UNICEF Venture Fund has been transformative. Beyond the financial investment, the team credits the program’s mentorship, technical guidance, and global exposure as key accelerators of their progress.