Exploring other alternatives for additional computational power
- Jan 1, 2026
- 2 min read
Dr Jacob Agyekum continued and completed the ethical clearance application process during this quarter. All required documentation, including the approved survey questionnaire and focus group discussion guides, was finalised and submitted through the online application platform. Delays in correspondence and administrative feedback have delayed the application process. Still, eventually approval was granted, enabling the study to proceed with data analysis and, subsequently, community engagement phases in compliance with ethical standards.
Preliminary statistical and exploratory analyses were conducted to examine the relationship between malaria incidence and key climate variables, including rainfall, temperature, and humidity. Climate and malaria datasets were temporally and spatially aligned to ensure consistency across sources. Initial trend analysis and correlation assessments were performed to identify seasonal patterns and potential lag effects between climate drivers and malaria outcomes. Managing inconsistencies across multiple data sources was challenging. Differences in temporal resolution, spatial coverage, and data completeness, notably missing values in climate datasets, required additional preprocessing and validation steps, increasing the time needed for analysis.
Initial machine learning model runs were conducted using the processed malaria and climate datasets. Data were structured into training and testing sets, and baseline models were implemented to assess predictive performance and identify influential variables. These initial runs provided valuable insights into model behaviour, data suitability, and areas requiring further refinement, forming the foundation for more advanced model optimisation in subsequent phases. The primary challenge in the initial machine learning model runs was addressing data imbalance and limited sample sizes in certain districts and time periods. This affected early model performance and required further model refinement. Additional computational power is needed, and Dr Agyekum is actively exploring other alternatives.
Plans for community engagement activities were developed in preparation for implementation in early February. The Health Director of Upper West Region in Ghana has been formally contacted and has expressed readiness to support and facilitate the community engagement process. This collaboration will support effective community participation and ensure that planned engagement activities align with local health priorities and implementation structures. The main challenge was coordinating schedules and aligning activities with health system workflows, despite the Health Directorate’s readiness to support the engagement.
At the same time, the Ghana Meteorological Agency supplied essential climate data. A social scientist supported the development of a survey instrument for community engagement. During data analysis, the data scientist supported dataset preparation, integrating complex climate, environmental, and health data to generate meaningful and accurate research outcomes.
Dr Jacob Agyekum reporting on his progress for the following research project: Harnessing Community Resilience and Machine Learning for Adaptive Malaria Control amid Climate Change in the Upper West Region of Ghana
Edited by Heidi Sonnekus & Leti Kleyn for the FAR-LeaF programme.
